2021 年 62 巻 8 号 p. 1102-1108
This paper describes the characterization of the interfacial properties of bonded dissimilar materials for the fabrication of biocompatible composites and functionally graded materials (FGMs) with high mechanical performance. In the development of biomaterials with conflicting properties such as high strength and toughness for artificial bone, much attention has been paid to composites and FGMs consisting of biocompatible ceramics and metals. Their mechanical properties are influenced by the properties of the interfaces of dissimilar materials, and hence it is important to evaluate the interfacial properties. This study investigated the influence of combinations of materials on the interfacial strength and fracture toughness of bonded dissimilar materials consisting of four types of biocompatible materials: titanium, type 316L stainless steel, partially stabilized zirconia, and alumina. The bonded dissimilar materials were fabricated via spark plasma sintering, which is a powder metallurgy technique utilizing uniaxial load and pulsed direct current in vacuum. The interfacial strength and toughness of the materials were evaluated via compression-bend testing and indentation testing, respectively. The distributions of elements near the interfaces due to atomic diffusion during sintering were evaluated, and the influence of material combinations on the interfacial properties was considered based on the distributions. It was found that the mechanical properties of all interfaces were lower than those of monolithic materials, and the extent of the degradation in mechanical properties was dependent on the material combinations. If atomic diffusion occurred on both sides of the interface, the interfacial fracture toughness and strength tended to be relatively high.
This Paper was Originally Published in Japanese in J. Soc. Mater. Sci., Japan 69 (2020) 855–862.
Biocompatible materials such as zirconia, alumina, and titanium have been used for implant applications to replace bones damaged by aging and injury.1–3) While the materials for such applications have required various mechanical characteristics with regard to toughness, strength, hardness, and wear resistance, none of the materials employed can satisfy all of the characteristics. Ceramics exhibit high wear resistance and hardness and low toughness, whereas metals exhibit low wear resistance and high toughness. Hence, much attention has been paid to the development of composites4–7) and functionally graded materials (FGMs)8–10) consisting of ceramics and metals, as it is expected that composites and FGMs will exhibit characteristics of both ceramics and metals. The authors have especially focused on FGMs in which the ceramic and metal phases are located on the surface and inside the FGMs, respectively, where the ceramic–metal composition gradually changes from the surface to the interior. Ceramic–metal composites11–14) and FGMs15,16) consisting of partially stabilized zirconia (PSZ), alumina (Al2O3), and pure titanium (Ti) were fabricated via spark plasma sintering,17) in which powders are sintered by pulsed currents with uniaxial pressure through graphite die and plungers. As a result, the fracture toughness of the PSZ–Ti and Al2O3–Ti composites was low and tended to decrease with increasing Ti volume fraction. Also, in both PSZ–Ti and Al2O3–Ti FGMs, cracks unstably grew from the ceramic surface to the metal substrate, indicating that the graded layers had little ability to arrest crack growth from the surface. The causes of the embrittlement in these composites were different: the PSZ–Ti composites exhibited brittleness due to the creation of brittle Ti oxide during sintering, while the Al2O3–Ti composites exhibited brittleness because fractures occurred along the weak interface between Al2O3 and Ti phases. Hence, it is important to clarify the formation of the interface of dissimilar materials and the effect of material combinations on the mechanical properties in developing biocompatible composites and FGMs with high mechanical characteristics.
The object of this study was to elucidate the formation of interfaces between dissimilar materials and the effect of material combinations on their mechanical properties. PSZ, Al2O3, Ti, and type 316L stainless steel (SUS316L, SUS) were used as representatives of biocompatible materials, and the behavior of the material combinations ceramic–metal,11–16) ceramic–ceramic,18) and metal–metal was examined. A bi-material specimen with a millimeter-sized interface of bonded dissimilar materials was fabricated for each combination of materials, and the strength and toughness of the interface were evaluated. In addition, the elemental concentration distributions in the vicinity of the interface were measured, and the formation of the interface was investigated. Then, the mechanical properties of the interface were investigated from the viewpoint of element distributions near the interface, and the development of composites and FGMs with high mechanical properties was considered.
Powders of Al2O3 (average particle size $\skew2\bar{d} = 27$ µm, AHP30, Nippon Light Metal Company, Ltd.), PSZ ($\skew2\bar{d} = 0.35$ µm, KZ-3YF-C, KCM Corporation), Ti ($\skew2\bar{d} = 23$ µm, TC-459, Toho Technical Service Co., Ltd.), SUS ($\skew2\bar{d} = 10$ µm, DAP316L, Daido Steel Co., Ltd.) were used. The chemical composition of these powders is listed in Table 1. Prior to sintering, each powder was milled in vacuum for 7 h using a vibration ball milling machine (NEVMA-8, Nissin Giken Co., Ltd.) to deagglomerate the as-received powders.
To evaluate the characteristics of the interfaces of bonded dissimilar materials, bi-material specimens with interfaces on a plane perpendicular to their longitudinal direction were fabricated, as shown in Fig. 1(a). As the sinterability of the constituent materials is different, the sintering conditions for each combination of materials were determined based on those for the material with the higher melting point, as listed in Table 2. Table 3 shows the sintering conditions for the combinations of materials. Note that this study did not aim for optimum sintering conditions. For Ti–SUS, a dense interface was successfully fabricated by sintering a layered arrangement of Ti and SUS powders, as shown in Fig. 1(b). On the other hand, the interfaces of Al2O3–Ti, PSZ–Ti, Al2O3–PSZ, and PSZ–SUS, fabricated with a similar one-step method, cracked after sintering due to residual stress induced by sintering, and these specimens could not be fabricated. Hence, a two-step sintering process was employed, as shown in Fig. 1(c): the material with the lower coefficient of thermal expansion was sintered first (pre-sintering), and then the powder of the other material was placed around it and was sintered under the same condition as for the pre-sintering. For Al2O3–SUS, no specimen could be fabricated because the SUS phase was melted at the temperature at which the densification of the Al2O3 phase started. Interfaces of SUS–Ti, Al2O3–Ti, PSZ–Ti, Al2O3–PSZ, and PSZ–SUS were examined in this study. A compact with each interface was sintered using a spark plasma sintering apparatus (SPS-211Lx, Fuji Electronic Industrial Co., Ltd.). During sintering, the temperature of the surface of the graphite die was measured using a pyrometer from outside the furnace, and was controlled. Figure 2 shows the changes in temperature and the displacement of the plunger during sintering (not pre-sintering) of the specimen with a PSZ–Ti interface. After reaching approximately 1100°C during heating, the displacement was constant, indicating that the densification was complete. In the sintering of all the combinations of materials, it was confirmed that the displacement became constant during heating and densification was completed, as shown in Fig. 2.
Schematic illustration of fabrication of pillar-shaped specimens with interface of dissimilar materials; (a) Dimensions of a specimen; (b) Sintering with one step; (c) Sintering with two steps.
Temperature and displacement measured during sintering of PSZ–Ti. Note that the temperature below 600°C could not be measured due to measurable range of the infrared pyrometer used.
Each specimen was cut from a sintered compact with a diamond saw, and all surfaces of the specimen were finished with sandpaper up to #800 and diamond lapping films up to #2000.
2.3 Evaluation of interfacial strength by compressive-bend testingBend testing was conducted for each combination of materials to evaluate the interfacial strength of the bi-material specimens. Interfacial strength can be evaluated using tensile testing and/or four-point bend testing;19–21) however it was difficult to apply such tests to our specimens due to the dimensional limitations of the sintered compacts fabricated using the SPS apparatus. Hence, we developed a compressive-bend testing technique in which a constant bend moment is applied to a small specimen. Figure 3 shows a schematic illustration of the testing procedure. Both edges of a specimen are fixed with flat jigs, and a compressive load P is applied to the specimen at a distance L from the specimen center. This subjects the specimen to bending and compression. The normal stress at the tensile side of the specimen (bending stress, σB) is given by the following equation:
| \begin{equation} \sigma_{\text{B}} = \frac{12PL}{bh^{3}}\cdot \frac{h}{2} - \frac{P}{bh}, \end{equation} | (1) |
Setup for compression-bending testing.
The evaluation technique for fracture toughness of an interface between a coating and substrate in thermal barrier coatings (TBCs) via indentation testing proposed by Lesage and Chicot22) was applied to the interfaces of bonded dissimilar materials. An interfacial crack was introduced by an indentation at the interface with a Vickers hardness tester, and the interfacial fracture toughness KC was calculated based on the length of the crack along the interface:
| \begin{equation} K_{\text{C}} = 0.015\frac{F}{a^{1.5}} \left[\frac{E}{H} \right]_{i}^{1/2}, \end{equation} | (2) |
| \begin{equation} \left[\frac{E}{H} \right]_{i}^{1/2} {}= \cfrac{\biggl(\cfrac{E}{H}\biggr)_{s}^{1/2}}{1+\biggl(\cfrac{H_{s}}{H_{c}}\biggr)^{1/2}} + \cfrac{\biggl(\cfrac{E}{H}\biggr)_{c}^{1/2}}{1+\biggl(\cfrac{H_{c}}{H_{s}}\biggr)^{1/2}}, \end{equation} | (3) |
For the interface of Ti–SUS, no interfacial crack could be introduced by the indentation technique. Hence, referring to the Japanese Industrial Standards (JIS) G0564, the interfacial fracture toughness was measured via three-point bend testing using a side-edge notched specimen. A notch with a radius of the notch root of 0.1 mm was created along the interface by electrical discharge machining. The testing was conducted in air at a crosshead speed of 0.05 mm/min. Assuming that the specimen was homogenous, the interfacial fracture toughness was calculated by
| \begin{equation} K_{\text{C}} = \left(\frac{FS}{bh^{3/2}} \right)\times G \left(\frac{d}{h}\right), \end{equation} | (4) |
| \begin{equation} G(\alpha) = 3\alpha^{\frac{1}{2}} \frac{1.99 - (\alpha) (1 - \alpha) (2.15 - 3.99\alpha + 2.7 \alpha^{2})}{2(1 + 2\alpha)(1 - \alpha)^{\frac{3}{2}}}, \end{equation} | (5) |
The elemental distributions near the interfaces of bonded dissimilar materials were measured using an electron probe microanalyzer (EPMA, JXA-8530F, JEOL). As will be described in Section 3.1, depending on the combination of materials, a layer of atomic diffusion due to sintering may occur near the interface (diffusion layer). Hence, the specimens were cut to expose the diffusion layer, and X-ray diffraction (XRD) patterns from the diffusion layer were obtained using an XRD system (RNT-2200, Rigaku) with a Cu target, acceleration voltage of 40 kV, and a current of 40 mA. The obtained XRD patterns were compared with the database PDF-2,23) and the reaction products created during sintering were identified.
Figure 4 shows the microstructure of the specimens observed by optical microscopy. The dissimilar materials are located at the top and bottom of the figure, and the vicinity of the center is the boundary. Observing their microstructures using OM, different materials were observed in different colors and contrast. As shown in the figure, discontinuous structural changes were confirmed at adjacent sections of the dissimilar materials, irrespective of the combination of materials. In this study, this discontinuous section was defined as the interface of the bonded dissimilar materials. For all material combinations, no large defects were observed near the interfaces. At the interfaces of PSZ–Ti and Ti–SUS, a gradual change in the microstructure from the base materials to the interfaces was observed on both sides of the interface, which is defined as the altered layer. The altered layer was observed only on the Ti side of the Al2O3–Ti interface. No altered layer was observed near the Al2O3–PSZ or PSZ–SUS interface.
Microstructure near the interfaces between dissimilar materials.
Figure 5(a) shows the distribution of elements near the PSZ–Ti interface where altered layers were observed on both sides. The distance from the interface was equal to the distance from the discontinuous section shown in Fig. 4. It was found that Ti diffused from the Ti side to the PSZ side and O diffused from the PSZ side to the Ti side. Figure 5(b) shows the XRD pattern in the region where O diffused in the Ti side. Ti2O and Ti2ZrO were detected, indicating that atoms diffused over the interface and formed reaction products. Figure 6 shows the PSZ–SUS interface where no altered layer was found. No atomic diffusion across the interface was found. A diffusion layer is defined as a layer in which more than 1 at% of the atoms in one material diffuses across the interface into the other material. Figure 7 shows the thickness of the diffusion layer near the interface for each combination of materials. The vertical axis denotes the distance from the interface, and the upper and lower sides of the figure show the materials corresponding to Materials A and B shown in Table 3, respectively. At the PSZ–Ti interface, diffusion layers of 50 µm and 200 µm thickness were formed on the PSZ and Ti sides, respectively. At the Al2O3–Ti interface, a diffusion layer of 350 µm thickness was formed only in the Ti side. No diffusion layer was observed near the PSZ–SUS and Al2O3–PSZ interfaces, while diffusion layers were formed in both sides near the Ti–SUS interface. These diffusion layers correspond to the altered layers in Fig. 4, and the formation of the diffusion layer is different for each combination of materials.
Interface between PSZ and Ti; (a) Distributions of Zr, O, and Ti; (b) XRD pattern.
Distributions of Zr, O, Fe, Cr, and Ni near the PSZ–SUS interface.
Thickness of diffusion layers.
Figure 8 shows typical bending stress–stroke diagrams. The stress–stroke relations were linear for all specimens, and brittle fracture occurred without large plastic deformation, irrespective of the combination of materials. Fractography indicated the presence of a debonding fracture at the interface between dissimilar materials.
Relationship between bending stress and stroke during compression-bending testing.
Figure 9 shows the relationship between the material combination and the interfacial strength. The scatter bars in the figure denote maximum and minimum measured values, and “No data” denotes that no measurement could be made due to specimen failure during processing. The ratio of the reduction of interfacial strength to the bending strength of the monolithic material is also shown. For ceramic–metal interfaces, the strength of the monolithic ceramic is used. For ceramic–ceramic and metal–metal interfaces, the strength of the material with the lower strength is used. For the strengths of the monolithic materials, the bending strengths obtained by the authors12,15) were used for monolithic ceramics, while the tensile strengths of the wrought materials24,25) were used for monolithic metals. The interfacial strength was highest for the Ti–SUS interface, and decreased in the order of PSZ–Ti and Al2O3–Ti. The strength of the interface of dissimilar metals was approximately 90% of that of the monolithic metal, while that of the ceramic–metal interfaces was less than approximately 30% of that of the monolithic materials. It should be noted that the stress singularity occurs at an edge between dissimilar materials, and it generally develops into a fracture.26) As all specimens were fractured from the interface edge, it would be appropriate to evaluate the strength considering their microstructure. However, depending on the combination of materials, a diffusion layer with a thickness of several hundreds of micrometers or more and reaction products (Fig. 5) were formed near the interface, and it is currently impossible to examine the stress singularities at interfaces with such complicated microstructures. In this study, as a first step for strength evaluation, each specimen with an interface is assumed to be homogeneous and isotropic.
Relationship between material combination and interfacial strength. The value of σmono denotes the strength of monolithic materials.
Figure 10 shows the relationship between the combination of materials and the interfacial fracture toughness. The rate of reduction in fracture toughness is also shown as in Fig. 9. The fracture toughness of PSZ–Ti and Al2O3–PSZ interfaces did not decrease significantly from that of monolithic PSZ and Al2O3, respectively, while that of the Al2O3–Ti and PSZ–SUS interfaces decreased significantly to less than approximately 30% of that of monolithic Al2O3 and PSZ, respectively.
Relationship between material combination and interfacial toughness. The scatter bars denote the maximum and minimum values, respectively, and KCmono denotes the fracture toughness of monolithic materials.
Table 4 summarizes the microstructure near the interfaces of bonded dissimilar materials and their mechanical properties. The reaction products identified using XRD are also shown in the table. The strength and fracture toughness of the interfaces were lower than those of monolithic materials in the scope of this study, irrespective of the combination of materials.
For ceramic–metal interfaces, the PSZ–Ti interface with diffusion layers on both sides of the interface exhibited the highest strength and fracture toughness, while the Al2O3–Ti interface with diffusion layers on only the Ti side and the PSZ–SUS interface with no diffusion layers exhibited significantly lower strength and fracture toughness. During sintering of bonded dissimilar materials, atomic diffusion occurs from one material into the other, and the sintering progresses by forming the interface between dissimilar materials. For combinations of materials for which diffusion layers were formed on both sides of the interface, atomic diffusion from one material to the other occurred actively, and a dense interface was formed. On the other hand, for combinations of materials for which no diffusion layer was formed, atomic diffusion was insufficient during sintering, and microscopic defects remained along the interface due to the interface being partially unsintered. As a result, the mechanical properties of the PSZ–Ti interface, in which diffusion layers were observed on both sides of the interface, tended to be superior to those of other material combinations. It is therefore necessary to select a combination of ceramic and metal materials that are prone to atomic diffusion and to optimize the sintering conditions to promote diffusion to fabricate a dense interface between dissimilar materials that exhibits high mechanical properties comparable to those of monolithic materials. It is known that atomic diffusion is strongly affected by the difference in crystal structure and atomic radius,27) and hence, atomistic evaluations including the interaction between atoms are necessary. The optical microscopy and the measurements of the element concentration distribution by EPMA conducted in this study were limited to investigations on a micrometer scale. It is possible to investigate the microstructural characteristics near the interfaces in more detail using ultra-high resolution scanning electron microscopy and/or transmission electron microscopy on the nanometer scale.
This study simply evaluated the interfacial strength assuming that the bi-material specimens are isotropic and homogeneous, as mentioned in Section 3.2. To properly evaluate the interfacial strength, it is necessary to consider the stress distribution near the interface and residual stress induced during sintering. Detailed observations of the microstructure and a mechanical evaluation considering the microstructure will help to clarify the strength development mechanism for ceramic–metal interfaces, and hence further study is necessary.
For the interface between dissimilar ceramics, although no diffusion layer was observed near the interface, the fracture toughness was about 50% of that of the monolithic ceramic (Al2O3), and the reduction in toughness was lower than that for other material combinations. For the interface between dissimilar metals, although the reduction in strength was not high compared to other combinations of materials, the interfacial fracture toughness was only 12% of that of monolithic metal (Ti), because brittle intermetallic compounds were formed at the interface. For the PSZ–Ti interface, intermetallic compounds and Ti oxides were formed as reaction products, but had only a slight effect on the reduction of interfacial fracture toughness. This is because the fracture toughness of PSZ is comparable to that of the reaction products. Therefore, the formation of brittle intermetallic compounds at the interface of dissimilar metals should be suppressed to fabricate an interface with high fracture toughness. In this study, only one interface was evaluated for dissimilar ceramics and dissimilar metals, respectively, and further investigations are needed to determine the relationship between material combinations and mechanical properties.
To clarify the effect of material combinations on the characteristics of the interface of bonded dissimilar materials, bi-material specimens with a millimeter-sized interface were fabricated for each combination of the materials PSZ, Al2O3, SUS, and Ti. The microstructure of these interfaces and their mechanical properties were investigated, and effective guidelines were provided for material selection in the development of composites and FGMs with excellent mechanical properties via SPS. The specific results obtained within the scope of this study are summarized below.
This work was supported by JSPS KAKENHI Grant Number 15H03891.
