Mechanical Properties Enhancement of Biomedical Au–Cu–Al Shape Memory Alloys by Phase Manipulation

Kang-Wei Goo; Wan-Ting Chiu; Akira Umise; Masaki Tahara; Masato Sone; Kenji Goto; Takao Hanawa; Hideki Hosoda

doi:10.2320/matertrans.MT-MBW2022001

Abstract

β-type Au–Cu–Al alloy, which is known as Spangold, has been developed for biomedical applications due to its high biocompatibility and X-ray contrast. Despite that, the workability of the polycrystalline functional phases (i.e., L2₁ parent and martensite phases) is limited by grain boundary embrittlement. To enhance the mechanical properties of the martensite phase (M-phase) alloy, this study aims to introduce a ductile α-fcc phase with various fractions into the M-phase alloy by manipulating the alloy chemical composition. Fundamental analysis such as phase identifications, thermal analysis, microstructure observations, and tensile tests were performed. As a result, the dual-phase M+α-fcc alloys with varying α-fcc phase fractions were successfully practiced. Among all alloys, their lattice parameters and phase transformation temperatures showed merely slight alteration, suggesting that the chemical compositions of the functional M-phases are similar for all alloys. Hence, the variations in the mechanical properties mainly originate from the different α-fcc phase fractions. The tensile test results indicate that both the ultimate tensile strength and fracture strain are promoted as the α-fcc phase fraction increases. The 50Au–38Cu–12Al alloy with the highest α-fcc phase fraction performs the most optimized mechanical properties.

1. Introduction

β-type gold–copper–aluminum (Au–Cu–Al) alloy, which is often called Spangold,¹^,²⁾ is highly biocompatible and the presence of heavy elements, such as Au, is preferred for medical applications due to its high X-ray contrast.³⁾ In addition, the shape memory effect (SME) was observed from the reversible phase transformation between the parent β-phase (L2₁ crystal structure)⁴⁾ and the martensite phase (M-phase, doubled B19 (DB19) crystal structure)⁵⁾ of these alloys. Hence, the functionality of SME further allows them to be a good candidate for biomedical applications.

However, the workability and applicability of the functional phases (M-phase and/or β-phase) in the polycrystalline Au–Cu–Al alloys are constrained by grain boundary embrittlement.⁶^,⁷⁾ Hence, the fracture strain of the functional phases was found to be lower than 11% in tensile tests⁸⁾ and 13% in compression tests.⁹⁾ It has been reported that the introduction of a ductile phase to the grain boundaries of a brittle phase, which is known as ductile phase toughening (DPT),¹⁰^,¹¹⁾ could strengthen the grain boundaries and increase the elongation of the alloys. Among the shape memory alloy community, strengthening of the brittle M-phase grain boundaries in the nickel–aluminum-based (Ni–Al-based) alloys, such as Ni–Al–Fe and Co–Ni–Al alloys, by introducing a ductile phase¹²^,¹³⁾ was successfully practiced. Besides that, similar advances were reported for the mechanical properties enhancement of the Au–Cu–Al-based alloys.¹⁴^–¹⁶⁾

This study, thus, aims to promote the mechanical properties of the M-phase alloy by introducing a ductile α-fcc phase (α-fcc, disordered fcc crystal structure)¹⁷⁾ in the Au–Cu–Al system. This study proves that both the strength and elongation of the alloys are promoted by introducing an appropriate fraction of the ductile α-fcc phase to the M-phase. To obtain alloys with various α-fcc phase fractions, the Au content was kept at 50 mol% as a constant, while the remaining elements were manipulated. The nominal compositions of the alloys used were 50Au–38Cu–12Al, 50Au–37Cu–13Al, and 50Au–36Cu–14Al (mol%), respectively. According to the 773 K isothermal ternary phase diagram of the Au–Cu–Al system,¹⁸⁾ these alloys would fall around the same tie-line; hence, the differences in the nominal composition would contribute to the variation in α-fcc phase fraction rather than affecting the chemical compositions of the M-phase and α-fcc phase.

2. Experimental Procedures

Au–Cu–Al alloys with nominal compositions of 50Au–38Cu–12Al, 50Au–37Cu–13Al, and 50Au–36Cu–14Al (mol%) were fabricated, respectively. Each ingot with a weight of 5 g was prepared by arc-melting pure Au, Cu, and Al metals (purity > 99.99%). The process was performed with a non-consumable tungsten electrode in a high-purity Ar–1 vol% H₂ atmosphere. The nominal chemical compositions, purity of elements, and abbreviations of the alloys are listed in Table 1. The abbreviations are used in the entire article unless otherwise mentioned.

Table 1 Nominal compositions, purity of elements, and abbreviations of the alloys.

The alloys were then processed into a disk shape by using a hot-forging at 873 K for 6 h in a high-purity Ar atmosphere. Thereafter, a homogenization treatment was carried out at 773 K for 1 h in a high-purity Ar atmosphere, followed by iced-water quenching.

Phase identification was analyzed on finely polished bulk samples by using an X-ray diffractometer (XRD; X’Pert PRO, PANalytical, UK) at room temperature (RT) of 296 K. The scanning range was from 20° to 120° with a scan rate of 0.042° s⁻¹ by using the CuK_α radiation (V = 40 kV, I = 45 mA, and λ = 0.15405 nm). A standard silicon plate was used as a reference for the correction of external errors. The CaRIne Crystallography software (Created by: Cyrille Boudias and Daniel Monceau, France) and CellCalc software (Created by Hiroyuki Miura, Japan) were used to determine the lattice parameters of the alloys. To determine the phase transformation temperature, a differential scanning calorimeter (DSC; DSC-60 Plus, Shimadzu, Japan) was conducted for 2 cycles and a high-purity Ar atmosphere was utilized to prevent the oxidation of specimens. The temperature range was from 223 K to 523 K with a scan rate of 10 K/min. The reference material used was alumina powder (Al₂O₃). A scanning electron microscope (SEM; S-4300SE, Hitachi, Japan) was utilized for microstructure observations. Mechanical properties were evaluated with a universal testing machine (Autograph AG-Xplus, Shimadzu, Japan) in a manner of tensile mode at RT of 296 K with a strain rate of 8.3 × 10⁻⁴ s⁻¹. The tensile specimen size was 10 mm in gauge length, 1 mm in width, and 0.3 mm in thickness, respectively.

3. Results and Discussion

The crystal structures of the phases in this alloy system are shown in Fig. 1. The α-fcc phase is with the fcc crystal structure and the M-phase possesses the doubled B19 crystal structure as mentioned in the introduction section. The lattice parameters indicated in Fig. 1 were calculated from the XRD results of alloy 12Al, while the parameters of other alloys are described in the following section. For the X-ray diffraction patterns of all alloys, please refer to Fig. 2.

Fig. 1

The crystal structures of the (a) α-fcc and (b) M-phases. Please note that the lattice parameters correspond to alloy 12Al.

Fig. 2

X-ray diffraction patterns of (a) 12Al, (b) 13Al, and (c) 14Al alloys at RT of 296 K. (The subscript of α represents the α-fcc phase).

The XRD patterns in Fig. 2 indicate that all alloys were composed of the dual-phase of the M+α-fcc at RT. Besides, the relative phase fraction could be approximated by comparing the peak intensities of both the α-fcc and M-phase. Here, the M-phase peak of 104_M and the α-fcc phase peak 111_α were used for the calculation of the peak ratios, which estimated the relative ratio of the α-fcc phase to the M-phase (α-fcc/M) as 0.85 for alloy 12Al, 0.75 for alloy 13Al, and 0.10 for alloy 14Al. It was found that the relative α-fcc phase fraction decreased in the sequence of alloy 12Al, 13Al, and 14Al (Fig. 2(a)–(c)). This trend is well consistent with the 773 K isothermal phase diagram of the Au–Cu–Al system.¹⁸⁾ Additionally, the M-phase lattice parameters of the alloys calculated were a = 0.450 nm, b = 0.587 nm, c = 1.787 nm for alloy 12Al, a = 0.450 nm, b = 0.587 nm, c = 1.782 nm for alloy 13Al, and a = 0.449 nm, b = 0.587 nm, c = 1.788 nm for alloy 14Al. The difference between the alloys is almost negligible. This signifies that the chemical compositions of the M-phase were similar among these 3 alloys. Further analysis concerning the similarity of the chemical composition of the M-phase is described below using the thermal analysis results.

The DSC curves displayed in Fig. 3 reveal that the austenite and martensite transformation peaks were clearly observed for all alloys. Furthermore, the area under the peaks signified that the relative M-phase fraction increases from alloy 12Al to alloy 14Al (Fig. 3(a)–(c)). On the contrary, it could be concluded that the relative α-fcc phase fraction decreases sequentially from alloy 12Al to alloy 14Al (Fig. 3(a)–(c)). This was in line with the findings from the XRD results (Fig. 2) and the assumption of the tie-line described in sections 1 and 2. In addition, the phase transformation temperatures indicated in Fig. 3 do not shift greatly when the Al content increased from alloy 12Al to alloy 14Al. This further proves the point mentioned in the XRD section, where the chemical compositions of the M-phase are similar among all three alloys. Thus, the chemical composition change of the alloys was mainly affecting the α-fcc phase fraction since these three alloys locate very close to the same tie-line. However, it should be noted that the precise and accurate chemical compositions of the M-phase and α-fcc phase could not be determined quantitively by using energy-dispersive X-ray spectroscopy (EDS) combined with SEM. This is because usually, the difference in the Al compositions between the M-phase and α-fcc phase is less than 1 mass%. It is also noted that the expected Al contents of equilibrated M- and α-fcc phases are approximately 3.2 mass% Al (in the alloy of 50 mol% Au–35 mol% Cu–15 mol% Al) and 2.3 mass% Al (in the alloy of 50 mol% Au–40 mol% Cu–10 mol% Al), respectively. It is also noted that the chemical compositions of the alloys could not be accurately determined by EDS area analysis since the difference of nominal Al compositions is only 0.4 mass% (2.57 mass% in alloy 12Al and 3.02 mass% in alloy 14Al). Nevertheless, again, from the experimental evidences of XRD and DSC, it could be determined that the alloys locate around the almost same tie-line.

Fig. 3

DSC results of (a) 12Al, (b) 13Al, and (c) 14Al alloys. (A_s, A_f, M_s, and M_f represent the austenite transformation start temperature, austenite transformation finish temperature, martensite transformation start temperature, and martensite transformation finish temperature, respectively.)

Figure 4 shows the microstructure of the alloys observed by SEM. First, the microstructure of the alloy 12Al (Fig. 4(a)), which was composed of the highest relative α-fcc phase fraction according to the X-ray diffraction pattern and thermal analysis curve, contained more α-fcc phase in the SEM image. Second, when the relative α-fcc phase fraction decreased (alloy 13Al), the amount of α-fcc phase decreased, as shown in Fig. 4(b). Lastly, when the relative α-fcc phase fraction was the lowest (alloy 14Al), only the M-phase with the typical microstructure of the martensite plates was observed. The presence of the α-fcc phase, which was observed by the X-ray diffraction measurements (Fig. 2(c)), could not be clearly identified since the relative amount of the α-fcc phase was low.

Fig. 4

SEM images of (a) 12Al, (b) 13Al, and (c) 14Al alloys, respectively.

The stress-strain curves (S-S curves) of all alloys examined by tensile tests at RT are shown in Fig. 5. By analyzing the results of Fig. 2, Fig. 3, and Fig. 5, the yield strength, ultimate tensile strength (σ_UTS), fracture strain (ε_f), and relative α-fcc phase fraction were summarized in Table 2. Both the σ_UTS and the ε_f of the alloys increase when the relative α-fcc phase fraction increases. Among these three alloys, the alloy 12Al, which possesses both the highest strength value and fracture strain value, was the most optimized one. As discussed above, the chemical compositions of the phases in all three alloys are almost identical. This suggested that the chemical composition effect on the mechanical properties could be neglected. In other words, the mechanical properties enhancement was contributed mainly by the increase in the α-fcc phase fraction. This proves that the addition of the α-fcc phase into the M-phase successfully improves the strength and elongation of the alloy 12Al. In contrast, alloys 13Al and 14Al indicate an insufficient α-fcc phase fraction since their fracture strains were lower than 11%, which is the elongation of the single M-phase alloy reported in the previous study.⁸⁾ Additionally, the S-S curve of alloy 12Al showed two-stage yielding, which could be attributed to the martensite variant reorientation under a plateau tensile stress. This proved that the appropriate amount of the α-fcc phase in alloy 12Al managed to enhance the mechanical properties while preserving the functionality of the M-phase. Whereas the microstructural control and the effects of the microstructures on the mechanical properties will be further investigated and developed in our future works.

Fig. 5

Stress-strain curves of (a) 12Al, (b) 13Al, and (c) 14Al alloys, which were examined by tensile tests at 296 K.

Table 2 Yield stress, ultimate tensile strength, fracture strain, and relative α-fcc phase fraction of (a) 12Al, (b) 13Al, and (c) 14Al alloys, respectively. The estimated relative phase fraction (α-fcc/M) was calculated from the relative intensity ratio between the 111_α and 104_M XRD peaks in Fig. 2.

A graph of σ_UTS-ε_f is shown in Fig. 6 to summarize the mechanical properties of the Au–Cu–Al based alloys. The solid symbols represent the results obtained from this study and the open symbols suggest the results from the previous study.¹⁶⁾ Basically, all the alloys are Au–Cu–Al M-phase and α-fcc dual-phase alloys. The main difference among the alloys is the phase fraction of these two phases as shown in the X-ray diffraction patterns (Fig. 2). Square symbols indicate the single α-fcc phase alloys, triangle symbols suggest the α-fcc phase-rich alloys, and circle symbols represent the M-phase-rich alloys. The results from the previous study showed that the introduction of the minor M-phase into the major α-fcc phase enhanced the mechanical properties of the alloys as well. However, compared to the previous study, alloy 12Al in this work demonstrates significantly optimized mechanical properties. Its mechanical properties are much higher than the single α-fcc phase alloy (square symbols) and the α-fcc phase-rich dual-phase alloys (triangle symbols). This concludes that the M-phase-rich alloy (circle symbols) with a relatively high α-fcc phase fraction is desired to achieve a good combination of high strength and elongation in the Au–Cu–Al alloys (i.e., the solid circle symbol at the upper-right corner).

Fig. 6

Ultimate tensile strength versus fracture strain plot of alloys composed of M-phase and/or α-fcc phase. (The solid circle symbols signify the results of this study while the open symbols indicate the results of the previous studies.¹⁶⁾)

4. Conclusions

The M+α-fcc dual-phase Au–Cu–Al alloys with various compositions were fabricated and the analysis of phase identifications, thermal analysis, microstructure observations, and mechanical property evaluations, were conducted. The important findings are listed in the following bullets:

(1) The 50Au–38Cu–12Al, 50Au–37Cu–13Al, and 50Au–36Cu–14Al dual-phase alloys, which were composed of the α-fcc and double B19 (DB19) martensite phase with various phase fraction ratios between these two phases, were attained via the manipulation of the nominal compositions.
(2) The relative α-fcc phase fraction increased from alloy 50Au–36Cu–14Al, 50Au–37Cu–13Al, and 50Au–38Cu–12Al sequentially.
(3) The mechanical properties of the Au–Cu–Al alloys were significantly improved by the introduction of the ductile α-fcc phase into the M-phase. The alloy with the highest α-fcc phase fraction (i.e., 50Au–38Cu–12Al) exhibited the best mechanical properties.
(4) The combination of major M-phase and minor α-fcc phase in the Au–Cu–Al alloys proved to be hopeful in mechanical properties enhancement; nonetheless, further studies should be made for controlling their microstructures.

Acknowledgments

This work is supported by the Japan Society for the Promotion of Science (JSPS) (KAKENHI 22K14491, 19H02417, 20K20544, 22H00256, and 21H01668). The authors would like to thank Ms. TORIYABE Ayano for the crystal structure data used in the XRD analysis.

REFERENCES

Corresponding author

Version information

Register with J-STAGE for free!