2019 Volume 60 Issue 5 Pages 662-665
The compressive deformation behavior and magnetic volume susceptibility were investigated for Au2CuAl biomedical shape memory alloys in a compositional range from 25 to 45 mol%Cu. Compression tests revealed that the stress for martensite variant reorientation was 282 MPa in Au2CuAl, and the value increased with the Cu content. On the other hand, the slip stress was higher at intermediate compositions. Moreover, intergranular fracture was suppressed during compressive deformation. Calculated antiphase boundary (APB) energies suggest the dissociation of superlattice dislocation, which leads to the active slip of ⟨111⟩-type. The measured magnetic volume susceptibility was −2.7 × 10−6 in Au2CuAl, hence, this alloy is judged to be metal-artifact-free in magnetic resonance imaging (MRI). The magnetic susceptibility increased up to +7.0 × 10−6 with increasing Cu content.
Intravascular treatment has attracted much attention recently. Intervention medical devices such as stents and guide wires are used under X-ray fluoroscopy.1) TiNi shape memory alloy (SMA) is often used in the intervention devices.2) However, concerns are raised on the possibility of Ni allergy and poor radio-opacity.2–4) Hence, a new biomedical SMA exhibiting good biocompatibility, shape memory properties and X-ray radiography is required. Based on the background, we have proposed Au2CuAl alloy as a promising candidate. Au2CuAl is composed of nontoxic elements only, contains heavy element Au, which allows high X-ray absorption, and possesses shape memory effect.5) A typical Au2CuAl alloy is expressed as Au7Cu5Al4 and called “Spangold” owing to its glittering character due to martensitic transformation and originally developed for jewelry.6,7) This alloy has A2 (bcc) - B2 - L21 (Heusler) order-disorder phase transformation,8,9) and the L21 parent phase causes martensitic transformation.10,11) The typical crystal structure of martensite is orthorhombic with c/a value of 0.943.11) Monoclinic phase is also reported in Au–18 mass%Cu–6 mass%Al alloy.10) On the contrary to the crystallographic information, limited data is available regarding mechanical properties such as Vickers hardness as a function of the Al content.12) Then, we have investigated the tensile mechanical properties of ternary AuCuAl alloys, and found that they are brittle when the Al content is higher than 15 mol%.13) In addition, the fracture mode is intergranular fracture, and Fe addition enhances ductility.14) However, there is still no classification on plastic deformation behavior of ternary Au2CuAl. Therefore, the first purpose of this study is to evaluate the deformation behavior by compression tests.
Magnetic resonance imaging (MRI) is a powerful observation tool; however, accurate diagnosis is often inhibited by metal artifacts. The metal artifacts is caused by difference in magnetization between organs and devices, which causes disturbance in the uniformity of the magnetic field.2) The metal artifacts in MRI can be suppressed by controlling magnetic volume susceptibility close to the human body (−9 ± 2 × 10−6)2) and several metal-artifacts-free materials have been developed such as AuPtNb,15) ZrNb,16) ZrMo17) and NbTaWZr.18) Then, the second purpose of this study is to clarify the magnetic susceptibility of Au2CuAl toward the intervention device for MRI.
The chemical compositions were Au–(25, 30, 35, 40, 45) mol%Cu–25 mol%Al. 10 g alloy ingots were produced by arc-melting with W electrode in Ar using 99.99% purity Au and Cu, and 99.9% Al. Since the maximum weight change after melting was 0.1 mass%, the nominal compositions were regarded as alloy composition. Hereafter, the alloys are named regarding the Cu content. For instance, 25Cu stands for Au–25 mol%Cu–25 mol%Al. The ingots were hot-pressed at 873 K in vacuum, and heat-treated at 773 K for 3.6 ks in Ar followed by water quenching (298 K). Then, disk-shape specimens with 2 mm of thickness were produced, and the test specimens were made by electrodischarge machining. For the phase identification, θ-2θ X-ray diffraction measurement (XRD: Philips X’pert Pro Galaxy System) was conducted at room temperature (RT, 298 K) using with CuKα radiation. Martensitic transformation temperatures were determined by differential scanning calorimetry (DSC: Bruker AXS DSC3200SA) in a temperature range from 223 to 423 K with a heating/cooling rate at 0.167 K/s. Compression tests were done by a tensile testing machine (Shimadzu Autograph AG-20kNXPlus) at RT with a constant strain rate of 3.3 × 10−4/s at RT (298 K). The specimen shape was a cylinder having 2 mm in diameter and 7 mm in height. Magnetic volume susceptibility was measured by a magnetic susceptibility balance (Sherwood Scientific LTD. MSB-AUTO) at RT (298 K) using rod shape specimens with 2 mm in diameter and 5 mm in height with hemisphere ends of 1R.
Figure 1 shows XRD profiles obtained at RT (298 K). The apparent phases evaluated from the XRD profiles are also listed in Table 1. All of the alloys except 45Cu contained L21 parent phase and orthorhombic martensite phase. Small peaks of monoclinic martensite phase were detected in 35Cu and 45Cu only. 45Cu was composed of the martensite phases only. Forward (M) and reverse (A) martensitic transformation temperatures (Ms, Mf, As and Af) were determined by DSC as shown in Fig. 2. The transformation temperatures were also listed in Table 1. The subscripts “s” and “f” stand for start and finish, respectively. In all of the alloys, only single peaks were detected, and all As were higher than RT regardless of the composition. In 25Cu, 30Cu and 35Cu, Ms were all lower than RT, but forward transformation was hardly detected in 40Cu and 45Cu, probably due to wide temperature hysteresis resulting to small temperature difference between samples and reference materials. It should be noted that the disappearance of forward martensitic transformation in DSC measurements has been sometimes reported in SMAs such as Nb-added AuTiCo.19) XRD results and DSC results are consistent with each other. Microstructure of 25Cu was reported.13)
Partial XRD profiles of Au2CuAl alloys. Open circles stand for the reflections from the parent phase (P), open and solid squares stand for the reflections from the orthorhombic (Mo) and the monoclinic (Mm) martensite phase, respectively.
Partial DSC curves of Au2CuAl alloys. Note that forward phase transformation was not detected in 40Cu and 45Cu.
Figure 3 shows stress-strain curves by the compression tests at 298 K. All of the alloys, except 40Cu, did not break during the compression but showed plastic deformation. Especially, 25Cu, 30Cu and 45Cu had good compressive deformability over 5%, and these results were quite different from those obtained by tensile tests.13) Therefore, it was revealed that (1) intergranular fracture is well suppressed during compression, and (2) Au2CuAl is not intrinsically brittle when grain boundary fracture is suppressed. Moreover, 25Cu, 30Cu and 45Cu all exhibited a two-stage yielding behavior. The two stage yielding behavior is often observed in shape memory alloys, and the first yielding corresponds to martensite variant reorientation (MVR) or stress-induced martensite formation. The second yielding often corresponds to dislocation slip (permanent deformation). On the other hand, 35Cu and 40Cu both exhibited single-stage yielding, and they buckled just after yielding. Then, the yielding behavior of 35Cu and 40Cu is judged to be slip deformation instead of MVR. This implies shape memory effect does not occur in these alloys. Figure 4 shows the results of MVR stress (σMVR) and slip stress (σSLIP) as a function of the Cu content. σMVR linearly increases with the Cu content. This is partially because As becomes higher with the Cu content. It is noted that σMVR in many SMAs is minimized at the temperature between Ms and As20) and that it increases with decreasing test temperature from martensitic transformation temperature such as NiTi, NiMnGa21) and (Au, Fe)Ti.22) On the other hand, σSLIP dose not simply increase with the Cu content but is higher at the intermediate composition, such as 35Cu. This may be partially due to solid solution (antisite defect) hardening by excess Cu substituting the Au sites.
Compressive stress-strain curves of Au2CuAl alloys.
Stress for martensite variant reorientation and slip stress evaluated in Au2CuAl alloys as a function of Cu content.
In order to evaluate the deformation behavior, antiphase boundary energy (APBE) was estimated. In the L21 type compounds, three possible slips are known: Burgers vector b = a⟨110⟩, b = 2a⟨100⟩, and b = 2a⟨111⟩, where a is the lattice constant of the bcc unit. It is commonly known from the von Mises criteria that bcc based intermetallics can be ductile when ⟨111⟩ dislocations move; at least five slip systems are active. For the activation of ⟨111⟩ slip, the perfect dislocation has to be dissociated into a combination of superpartial dislocations with APBs, and this must occur when APBE is sufficiently low. The dislocation dissociation of L21 compounds is described as: a/2⟨111⟩ / APB(I) / a/2⟨111⟩ /ABP(II) / a/2⟨111⟩/ APB(I) / a/2⟨111⟩.23) Then, APBEs on {100}, {110}, {112} and {123} of 25Cu (stoichiometry) were evaluated based on APBE calculation for B2 compounds24) and listed in Table 2. The assumptions are: the long range order parameter of 1, the interaction energies between unlike atoms9) listed in Table 3, and the experimental lattice parameter (2a = 0.623 nm). The distance between separated superpartial screw dislocations (= width of APB) can be estimated by the equation and also listed in Table 2;
| \begin{equation} r = G\boldsymbol{{b}}^{2}/2\pi\gamma_{\text{APB}}, \end{equation} | (1) |
Figure 5 shows the magnetic volume susceptibility as a function of the Cu content. The magnetic susceptibility was ranged from −2.7 to +7 (×10−6). The result revealed that the absolute values of magnetic volume susceptibility of Au2CuAl alloys are quite small in comparison with other nonmagnetic metals (e.g.: Ti = 182, Cr = 320, Zr = 109 and Pt = 279 (×10−6)),26) and that the stoichiometric Au2CuAl (25Cu) alloy has similar magnetic susceptibility to that of human body (−9 ± 2 × 10−6). In this case, Au2CuAl alloy is judged to be MRI metal-artifact-free. A real observation of MRI is now planned.
As for the composition dependence, the magnetic susceptibility increases with increasing Cu content. In order to estimate the compositional effect, the volume magnetic susceptibility χvt is calculated by the following equation,27)
| \begin{equation} \chi_{vt} = d_{t}\left\{\left(\frac{W_{1}}{W_{t}}\cdot\frac{\chi_{v1}}{d_{1}}\right) + \left(\frac{W_{2}}{W_{t}}\cdot \frac{\chi_{v2}}{d_{2}}\right) + \cdots \left(\frac{W_{n}}{W_{t}}\cdot \frac{\chi_{vn}}{d_{n}}\right)\right\}, \end{equation} | (2) |
This work was supported by Grant-in-Aid for Scientific Research Kiban S 26220907 from Japan Society for the Promotion of Science (JSPS). We would like to thank Prof. Tso-Fu Mark Chang for his support to complete this research.
