Microstructure Evolution and Mechanical Properties of Titanium / Alumina Brazed Joints for Medical Implants

: Medical titanium and alumina (Al 2 O 3 ) bioceramic are widely utilized as biomaterials. A reliable brazed joint of titanium and alumina was successfully obtained using biocompatible Au foil for implantable devices in the present study. The interfacial microstructure and reaction products of titanium / Au / Al 2 O 3 joints brazed under di ﬀ erent conditions were investigated by scanning electron microscopy (SEM), energy dispersive spectroscopy (EDS), and X-ray di ﬀ raction (XRD). In this study, the typical interfacial microstructure of the titanium / Au / Al 2 O 3 joint was titanium / Ti 3 Au layer / TiAu layer / TiAu 2 layer / TiAu 4 layer / Au + granular TiAu 4 layer / TiO x phase / Al 2 O 3 ceramic. With increasing brazing temperature or holding time, the thicknesses of Ti 3 Au + TiAu + TiAu 2 layers adjacent to the titanium substrate increased gradually. Shear tests indicated that the joint brazed at 1115 ◦ C for 3 min exhibited the highest shear strength of 39.2 MPa. Typical fracture analysis displayed that the crack started at the Al 2 O 3 ceramic and propagated along the interface of TiAu 2 and TiAu 4 reaction layers.


Introduction
Titanium and its alloys have been intensively investigated and applied for biomedical applications because of their excellent biocompatibilities, mechanical properties, and corrosion resistances [1][2][3][4][5][6]. Applications have included dental implants, craniomaxillofacial implants, implants for artificial joint replacement and spinal components, internal fixation plates and screws, and housings for ventricular-assist devices and pacemaker cases [7][8][9]. Alumina, a ceramic with outstanding physical, chemical, and mechanical performances, has attracted great interest in industrial applications such as biomaterials, aerospace, nuclear power, automobiles, and electronics [10][11][12][13]. With excellent advantages in chemical stability, wear resistance, and biocompatibility, alumina has been a preferable orthopedic implant material used in dental and bone replacements as well as coatings for metallic materials [9,[14][15][16]. Utilization of metal-ceramic composites for biomedical applications, including implantable pacemakers, retinal implants, and microstimulators, has dramatically increased in recent To obtain titanium/Au/Al2O3 brazed joints, the joining surface of titanium was ground to a grit of 3000 with emery paper. The Al2O3 ceramic, Au foil, and α-titanium were all cleaned with acetone in an ultrasonic bath for 15 min, and then they were assembled as a sandwich structure, as described in Figure 1c. The atmosphere of the vacuum furnace was maintained at 1.3 × 10 −3 Pa during the brazing process. The furnace was firstly heated to 1000 °C for 10 min at a rate of 20 °C/min then to the brazing temperatures at a rate of 10 °C/min. Afterwards, in order to investigate the impact of brazing temperature on the microstructures and mechanical properties of the brazed joints, the brazing specimens were held for 1 min at different brazing temperatures. Finally, the specimens were cooled down to 300 °C at a rate of 5 °C/min and then to room temperature spontaneously in the furnace. To investigate the effect of holding time on the microstructures and mechanical properties of the brazed joints, the brazing specimens were kept for different holding times at 1115 °C. About 10 specimens were prepared in the same condition for each parameter.
The cross-sections of titanium/Au/Al2O3 brazed joints were characterized using SEM (MERLIN Compact, ZEISS, Stuttgart, Germany) equipped with an energy-dispersive X-ray spectrometer (EDS, Octane Plus, EDAX, Mahwah, NJ, USA). Shear tests were conducted on at least six specimens at room temperature at a constant strain rate of 1 mm/min using a universal testing machine (Instron 5967, Instron, Boston, MA, USA). Average values and deviations of shear strengths were calculated from five specimens after removing outliers for each parameter. After the shear test, the fractures of titanium/Au/Al2O3 brazed joints were analyzed by three fractured specimens selected randomly using SEM and X-ray diffraction (XRD, JDX-3530M). To further evaluate mechanical properties of the reaction products in the joint, the hardness and elastic modulus across the joints were measured using a nanoindenter (G200, Agilent, Santa Clara, CA, USA).

Typical Interfacial Microstructure of the Titanium/Au/Al2O3 Joint
Vacuum brazing of titanium alloy and Al2O3 ceramic was achieved using Au filler foil. Figure 2 shows the typical interfacial microstructure and the main element distribution of the titanium/Au/Al2O3 joint brazed at 1115 °C for 1 min. As shown in Figure 2a, according to the different contrasts of each phase by SEM in the backscattered electron mode, the joint could be classified into four continuous reaction zones (zone I to IV), and zone V adjacent to the Al2O3 substrate consisted of a white phase dispersed with some light grey granular phases. Figure 2a also shows variations of elemental concentration of Ti, Au, Al, and O along the white dashed line. The concentration profile of element Ti showed a stepwise decrease from titanium substrate to Al2O3 ceramic, with a noticeable enrichment on the metal/Al2O3 interface. Meanwhile, the main distribution of element Au in the seam exhibited a stepwise increase from titanium to Al2O3 ceramic, and it displayed a minute amount in titanium substrate. The elements of Al and O were mainly distributed in Al2O3 ceramic. These results were consistent with the corresponding elemental distribution in the typical joint displayed in Figure 2b-e.
The interdiffusion of Ti and Au indicated that the active element Ti diffused from the titanium substrate and spread in the brazing seam, and it eventually accumulated adjacent to the surface of To obtain titanium/Au/Al 2 O 3 brazed joints, the joining surface of titanium was ground to a grit of 3000 with emery paper. The Al 2 O 3 ceramic, Au foil, and α-titanium were all cleaned with acetone in an ultrasonic bath for 15 min, and then they were assembled as a sandwich structure, as described in Figure 1c. The atmosphere of the vacuum furnace was maintained at 1.3 × 10 −3 Pa during the brazing process. The furnace was firstly heated to 1000 • C for 10 min at a rate of 20 • C/min then to the brazing temperatures at a rate of 10 • C/min. Afterwards, in order to investigate the impact of brazing temperature on the microstructures and mechanical properties of the brazed joints, the brazing specimens were held for 1 min at different brazing temperatures. Finally, the specimens were cooled down to 300 • C at a rate of 5 • C/min and then to room temperature spontaneously in the furnace. To investigate the effect of holding time on the microstructures and mechanical properties of the brazed joints, the brazing specimens were kept for different holding times at 1115 • C. About 10 specimens were prepared in the same condition for each parameter.
The cross-sections of titanium/Au/Al 2 O 3 brazed joints were characterized using SEM (MERLIN Compact, ZEISS, Stuttgart, Germany) equipped with an energy-dispersive X-ray spectrometer (EDS, Octane Plus, EDAX, Mahwah, NJ, USA). Shear tests were conducted on at least six specimens at room temperature at a constant strain rate of 1 mm/min using a universal testing machine (Instron 5967, Instron, Boston, MA, USA). Average values and deviations of shear strengths were calculated from five specimens after removing outliers for each parameter. After the shear test, the fractures of titanium/Au/Al 2 O 3 brazed joints were analyzed by three fractured specimens selected randomly using SEM and X-ray diffraction (XRD, JDX-3530M). To further evaluate mechanical properties of the reaction products in the joint, the hardness and elastic modulus across the joints were measured using a nanoindenter (G200, Agilent, Santa Clara, CA, USA).

Typical Interfacial Microstructure of the Titanium/Au/Al 2 O 3 Joint
Vacuum brazing of titanium alloy and Al 2 O 3 ceramic was achieved using Au filler foil. Figure 2 shows the typical interfacial microstructure and the main element distribution of the titanium/Au/Al 2 O 3 joint brazed at 1115 • C for 1 min. As shown in Figure 2a, according to the different contrasts of each phase by SEM in the backscattered electron mode, the joint could be classified into four continuous reaction zones (zone I to IV), and zone V adjacent to the Al 2 O 3 substrate consisted of a white phase dispersed with some light grey granular phases.  In order to investigate microstructure characteristics of the titanium/Au/Al2O3 joint in detail, Figure 3 shows the microstructures of zones I-V under high magnification. EDS data showing elemental concentrations and possible phases at each spot are listed in Table 1. The characteristic areas of I to VI adjacent to titanium are shown in Figure 3a under high magnification. The characteristic areas of VI and V next to Al2O3 ceramic are displayed in Figure 3b under high magnification. According to the EDS results shown in Table 1 and the Ti-Au binary phase diagram illustrated in Figure 4 [41], the reaction layers that formed from the titanium substrate to Al2O3 substrate were a Ti3Au phase (Spot A), a TiAu phase (Spot B), a TiAu2 phase (Spot C), a TiAu4 phase (Spot D), and an Au phase (Spot E) containing TiAu4 particles, respectively.
From the above analysis, it was proposed that during brazing, active element Ti spread and accumulated on the metal/Al2O3 ceramic interface, which could be deduced from the reaction with Al2O3 and the formation of TiOx [25,30,31,42,43]. However, TiOx was hard to observe with SEM and EDS owing to its limited thickness. Apart from reacting with Al2O3, the dissolved Ti in the brazing seam also reacted with molten Au, forming Ti-Au intermetallic compounds (IMCs).
To sum up, the typical interfacial microstructure of the titanium/Au/Al2O3 joint brazed at 1115 °C for 1 min consisted of titanium/Ti3Au layer/TiAu layer/TiAu2 layer/TiAu4 layer/Au + granular TiAu4 layer/TiOx phase/Al2O3 ceramic.  In order to investigate microstructure characteristics of the titanium/Au/Al 2 O 3 joint in detail, Figure 3 shows the microstructures of zones I-V under high magnification. EDS data showing elemental concentrations and possible phases at each spot are listed in Table 1. The characteristic areas of I to VI adjacent to titanium are shown in Figure 3a under high magnification. The characteristic areas of VI and V next to Al 2 O 3 ceramic are displayed in Figure 3b under high magnification. According to the EDS results shown in Table 1 and the Ti-Au binary phase diagram illustrated in Figure 4 [41], the reaction layers that formed from the titanium substrate to Al 2 O 3 substrate were a Ti 3 Au phase (Spot A), a TiAu phase (Spot B), a TiAu 2 phase (Spot C), a TiAu 4 phase (Spot D), and an Au phase (Spot E) containing TiAu 4 particles, respectively.
From the above analysis, it was proposed that during brazing, active element Ti spread and accumulated on the metal/Al 2 O 3 ceramic interface, which could be deduced from the reaction with Al 2 O 3 and the formation of TiO x [25,30,31,42,43]. However, TiO x was hard to observe with SEM and EDS owing to its limited thickness. Apart from reacting with Al 2 O 3 , the dissolved Ti in the brazing seam also reacted with molten Au, forming Ti-Au intermetallic compounds (IMCs).
To sum up, the typical interfacial microstructure of the titanium/Au/Al 2 O 3 joint brazed at 1115 • C for 1 min consisted of titanium/Ti 3 Au layer/TiAu layer/TiAu 2 layer/TiAu 4 layer/Au + granular TiAu 4 layer/TiO x phase/Al 2 O 3 ceramic.  In order to investigate microstructure characteristics of the titanium/Au/Al2O3 joint in detail, Figure 3 shows the microstructures of zones I-V under high magnification. EDS data showing elemental concentrations and possible phases at each spot are listed in Table 1. The characteristic areas of I to VI adjacent to titanium are shown in Figure 3a under high magnification. The characteristic areas of VI and V next to Al2O3 ceramic are displayed in Figure 3b under high magnification. According to the EDS results shown in Table 1 and the Ti-Au binary phase diagram illustrated in Figure 4 [41], the reaction layers that formed from the titanium substrate to Al2O3 substrate were a Ti3Au phase (Spot A), a TiAu phase (Spot B), a TiAu2 phase (Spot C), a TiAu4 phase (Spot D), and an Au phase (Spot E) containing TiAu4 particles, respectively.
From the above analysis, it was proposed that during brazing, active element Ti spread and accumulated on the metal/Al2O3 ceramic interface, which could be deduced from the reaction with Al2O3 and the formation of TiOx [25,30,31,42,43]. However, TiOx was hard to observe with SEM and EDS owing to its limited thickness. Apart from reacting with Al2O3, the dissolved Ti in the brazing seam also reacted with molten Au, forming Ti-Au intermetallic compounds (IMCs).
To sum up, the typical interfacial microstructure of the titanium/Au/Al2O3 joint brazed at 1115 °C for 1 min consisted of titanium/Ti3Au layer/TiAu layer/TiAu2 layer/TiAu4 layer/Au + granular TiAu4 layer/TiOx phase/Al2O3 ceramic.   In order to illuminate the formation mechanism of the typical interfacial microstructure and different Ti-Au IMCs in the titanium/Au/Al 2 O 3 joint, the Ti-Au binary system was studied using the phase diagram shown in Figure 4 [41]. The complex interfacial microstructural morphology and arrangement of various intermetallic compounds (IMCs) generated during the brazing process were strongly dependent on the brazing temperature. The brazing process of titanium to Al 2 O 3 ceramic can be deduced as follows. According to the Ti-Au binary phase diagram (Figure 4), it was observed that element Au melted to the liquid phase when the temperature exceeded 1064 • C. The active element Ti diffused from titanium substrate and dissolved into liquid Au gradually. As shown in Figure 4, marked by the red line at 1115 • C, with an increasing Ti concentration in the liquid phase, Ti began to react with molten Au to form TiAu 4 IMC by the peritectic reaction of Au(L) + Ti → TiAu 4 . Thus, a continuous TiAu 4 layer formed in the brazing seam and inhibited the interdiffusion of Ti and Au. Because of the decreasing concentration gradient of Ti from titanium to the TiAu 4 layer, the Ti 3 Au, TiAu, and TiAu 2 layers simultaneously formed between titanium and the TiAu 4 layer. During the cooling process, TiAu 4 particles and the Au phase directly precipitated in the remnant liquid phase by the eutectic reaction of L → Au + TiAu 4 adjacent to the ceramic substrate. Finally, the typical microstructure of titanium/Au/Al 2 O 3 joint with five reaction zones was obtained in the brazing seam, as illustrated in Figure 3.  In order to illuminate the formation mechanism of the typical interfacial microstructure and different Ti-Au IMCs in the titanium/Au/Al2O3 joint, the Ti-Au binary system was studied using the phase diagram shown in Figure 4 [41]. The complex interfacial microstructural morphology and arrangement of various intermetallic compounds (IMCs) generated during the brazing process were strongly dependent on the brazing temperature. The brazing process of titanium to Al2O3 ceramic can be deduced as follows. According to the Ti-Au binary phase diagram (Figure 4), it was observed that element Au melted to the liquid phase when the temperature exceeded 1064 °C. The active element Ti diffused from titanium substrate and dissolved into liquid Au gradually. As shown in Figure 4, marked by the red line at 1115 °C, with an increasing Ti concentration in the liquid phase, Ti began to react with molten Au to form TiAu4 IMC by the peritectic reaction of Au(L) + Ti → TiAu4. Thus, a continuous TiAu4 layer formed in the brazing seam and inhibited the interdiffusion of Ti and Au. Because of the decreasing concentration gradient of Ti from titanium to the TiAu4 layer, the Ti3Au, TiAu, and TiAu2 layers simultaneously formed between titanium and the TiAu4 layer. During the cooling process, TiAu4 particles and the Au phase directly precipitated in the remnant liquid phase by the eutectic reaction of L → Au + TiAu4 adjacent to the ceramic substrate. Finally, the typical microstructure of titanium/Au/Al2O3 joint with five reaction zones was obtained in the brazing seam, as illustrated in Figure 3.  Figure 5 displays the microstructure evolution of the joints brazed at different temperatures for 1 min. Brazing temperatures had a significant effect on the interfacial microstructure, and the thicknesses of reaction layers were measured and illustrated in Figure 5f. With increasing temperature, the thicknesses of Ti3Au + TiAu + TiAu2 layers (zone I-III) adjacent to the titanium substrate increased gradually (Figure 5a-e). The thickness of the TiAu2 layer (zone III) increased first and then decreased, and the maximum thickness of 17.4 μm was obtained under 1115 °C. Meanwhile, as the brazing temperature increased, the thickness of the Au layer with granular TiAu4 (zone V) next to Al2O3 ceramic notably decreased.    Figure 6 shows the microstructure evolution of the joints brazed at 1115 °C for different holding times. The microstructure of the joints changed significantly with the prolongation of holding time, and the thickness of the reaction layers were measured and illustrated in Figure 6d. As shown in Figure 6a-c, with the prolongation of holding time from 1 to 5 min, the thicknesses of Ti3Au + TiAu + TiAu2 layers (zone I-III) increased. The thickness of the TiAu2 layer (zone III) increased first and then decreased, and the maximum thickness was obtained for a holding time of 3 min. Meanwhile, as the holding time increased, the thickness of the Au layer with granular TiAu4 (zone V) next to Al2O3 ceramic did not change significantly. Based on the above analyses, brazing temperature and holding time, which affected the dissolution of Ti from titanium substrate, had significant effects on the microstructure evolution of the joints. A conceptual model was established and illustrated in Figure 7 to show the evolution of the microstructure. The reaction process could be classified into three stages. As shown in Figure 7a, during the brazing process when the temperature was above the melting point of Au, Au foil first converted into liquid. Then, Ti dissolved into molten Au under the driving force of the concentration   Figure 6 shows the microstructure evolution of the joints brazed at 1115 °C for different holding times. The microstructure of the joints changed significantly with the prolongation of holding time, and the thickness of the reaction layers were measured and illustrated in Figure 6d. As shown in Figure 6a-c, with the prolongation of holding time from 1 to 5 min, the thicknesses of Ti3Au + TiAu + TiAu2 layers (zone I-III) increased. The thickness of the TiAu2 layer (zone III) increased first and then decreased, and the maximum thickness was obtained for a holding time of 3 min. Meanwhile, as the holding time increased, the thickness of the Au layer with granular TiAu4 (zone V) next to Al2O3 ceramic did not change significantly. Based on the above analyses, brazing temperature and holding time, which affected the dissolution of Ti from titanium substrate, had significant effects on the microstructure evolution of the joints. A conceptual model was established and illustrated in Figure 7 to show the evolution of the microstructure. The reaction process could be classified into three stages. As shown in Figure 7a, during the brazing process when the temperature was above the melting point of Au, Au foil first converted into liquid. Then, Ti dissolved into molten Au under the driving force of the concentration Based on the above analyses, brazing temperature and holding time, which affected the dissolution of Ti from titanium substrate, had significant effects on the microstructure evolution of the joints. Figure 7 to show the evolution of the microstructure. The reaction process could be classified into three stages. As shown in Figure 7a, during the brazing process when the temperature was above the melting point of Au, Au foil first converted into liquid. Then, Ti dissolved into molten Au under the driving force of the concentration gradient, and it reacted with Au to form the TiAu 4 layer between the Ti substrate and Au (Figure 7b). Finally, the Ti 3 Au, TiAu, and TiAu 2 layers simultaneously formed between titanium and the TiAu 4 layer along the concentration gradient of Ti. During the cooling process, TiAu 4 particles and the Au phase directly precipitated because the residual element Ti was present in the remnant liquid phase adjacent to the ceramic substrate (Figure 7c). When brazing temperature or holding time increased, the mutual diffusion of Ti and Au became more sufficient. As a result, the thicknesses of Ti 3 Au + TiAu + TiAu 2 layers increased gradually, especially the TiAu 2 layer. Meanwhile, the Au phase containing TiAu 4 particles reduced, as shown in Figure 7d. With the further increase of brazing temperature or holding time, the diffusion of Au was adequate, and the mount of Ti was sufficient. Ti 3 Au and TiAu layers increased, resulting in the decreased thickness of the TiAu 2 layer. It was notable that the TiAu 4 layer almost occupied the brazing seam next to the ceramic (Figure 7e).

A conceptual model was established and illustrated in
It has been widely reported that TiO x could be generated on the interface of Ti containing metal and Al 2 O 3 [25,30,32,35,[38][39][40]. The limited thickness of TiO x and many other compounds in metal-ceramic interfaces led to a decreased accuracy in the identification of titanium oxides [20,44]. As brazing temperature or holding time rose, more TiO x phases formed adjacent to Al 2 O 3 (Figure 7c-e).
Metals 2019, 9, x FOR PEER REVIEW 7 of 12 gradient, and it reacted with Au to form the TiAu4 layer between the Ti substrate and Au (Figure 7b). Finally, the Ti3Au, TiAu, and TiAu2 layers simultaneously formed between titanium and the TiAu4 layer along the concentration gradient of Ti. During the cooling process, TiAu4 particles and the Au phase directly precipitated because the residual element Ti was present in the remnant liquid phase adjacent to the ceramic substrate (Figure 7c). When brazing temperature or holding time increased, the mutual diffusion of Ti and Au became more sufficient. As a result, the thicknesses of Ti3Au + TiAu + TiAu2 layers increased gradually, especially the TiAu2 layer. Meanwhile, the Au phase containing TiAu4 particles reduced, as shown in Figure 7d. With the further increase of brazing temperature or holding time, the diffusion of Au was adequate, and the mount of Ti was sufficient. Ti3Au and TiAu layers increased, resulting in the decreased thickness of the TiAu2 layer. It was notable that the TiAu4 layer almost occupied the brazing seam next to the ceramic (Figure 7e). It has been widely reported that TiOx could be generated on the interface of Ti containing metal and Al2O3 [25,30,32,35,[38][39][40]. The limited thickness of TiOx and many other compounds in metalceramic interfaces led to a decreased accuracy in the identification of titanium oxides [20,44]. As brazing temperature or holding time rose, more TiOx phases formed adjacent to Al2O3 (Figure 7c-e).

Mechanical Properties and Fracture Morphology of Titanium/Au/Al2O3 Joints
In order to evaluate the effect of brazing temperature and holding time on the mechanical properties of brazed joints, the shear strength of the joints was tested at room temperature, as shown in Figure 8. As shown in Figure 8a, when the joints were brazed at different temperatures varying from 1105 to 1125 °C for 1 min, the shear strength of the joints increased first and then decreased. The maximum average shear strength of 20.3 MPa was obtained when the joints were brazed at 1105 °C for 1 min.
As shown in Figure 8b, the shear strength of the joints firstly increased and then decreased when the joints were brazed at 1115 °C for different holding times prolonged from 1 to 5 min. The maximum value of shear strength reached 39.2 MPa when the holding time was 3 min, which was about twice that of the joints brazed at 1115 °C for 1 min.

Mechanical Properties and Fracture Morphology of Titanium/Au/Al 2 O 3 Joints
In order to evaluate the effect of brazing temperature and holding time on the mechanical properties of brazed joints, the shear strength of the joints was tested at room temperature, as shown in Figure 8. As shown in Figure 8a, when the joints were brazed at different temperatures varying from 1105 to 1125 • C for 1 min, the shear strength of the joints increased first and then decreased. The maximum average shear strength of 20.3 MPa was obtained when the joints were brazed at 1105 • C for 1 min.
As shown in Figure 8b, the shear strength of the joints firstly increased and then decreased when the joints were brazed at 1115 • C for different holding times prolonged from 1 to 5 min. The maximum value of shear strength reached 39.2 MPa when the holding time was 3 min, which was about twice that of the joints brazed at 1115 • C for 1 min. Fracture analysis was conducted using an optical microscope, SEM, and XRD to investigate the fracture location and fracture path of the titanium/Au/Al2O3 joints brazed at 1115 °C for 1 min. As shown in Figure 9a,b, Al2O3 ceramic was observed on the fracture surface of the titanium side. Figure  9d shows the crack was initiated at the Al2O3 ceramic and propagated into the brazing seam via the Au/Al2O3 interface during the shear test. The magnified SEM image of Figure 9d is shown in Figure  9e. When the crack propagated into the brazing seam, the joints fractured along the interface of TiAu2 and TiAu4 reaction layers (the interface of zone III and zone IV). The joints brazed at 1115 °C for 1 min fractured in the brittle mode. To further investigate the fracture location, reaction phases on the fracture surface of the titanium side were identified using XRD analysis, as shown in Figure 9c. It was evident that the fracture surface of the titanium side consisted of Au, Al2O3, and TiAu2, which corresponded to the fracture path analyses of Figure 9d,e.  Figure 10 displays the fracture analyses of the joints brazed at different parameters. It was observed that two types of fracture patterns existed after the shear test. In the first fracture pattern, significantly flat fracture surfaces were clearly observed, and the joints fractured along the Au/Al2O3 interface during the shear tests when brazed at 1105 °C for 1 min and 1115 °C for 5 min (Figures 10a  and 9c). XRD analyses of the fracture surface on the titanium side displayed detectable phases, including Au and Al2O3, which in turn supported the above analysis of the first fracture pattern. Meanwhile, as shown in Figure 10b, a second type of fracture pattern was observed when the joints Fracture analysis was conducted using an optical microscope, SEM, and XRD to investigate the fracture location and fracture path of the titanium/Au/Al 2 O 3 joints brazed at 1115 • C for 1 min. As shown in Figure 9a,b, Al 2 O 3 ceramic was observed on the fracture surface of the titanium side. Figure 9d shows the crack was initiated at the Al 2 O 3 ceramic and propagated into the brazing seam via the Au/Al 2 O 3 interface during the shear test. The magnified SEM image of Figure 9d is shown in Figure 9e. When the crack propagated into the brazing seam, the joints fractured along the interface of TiAu 2 and TiAu 4 reaction layers (the interface of zone III and zone IV). The joints brazed at 1115 • C for 1 min fractured in the brittle mode. To further investigate the fracture location, reaction phases on the fracture surface of the titanium side were identified using XRD analysis, as shown in Figure 9c. It was evident that the fracture surface of the titanium side consisted of Au, Al 2 O 3 , and TiAu 2 , which corresponded to the fracture path analyses of Figure 9d,e. Fracture analysis was conducted using an optical microscope, SEM, and XRD to investigate the fracture location and fracture path of the titanium/Au/Al2O3 joints brazed at 1115 °C for 1 min. As shown in Figure 9a,b, Al2O3 ceramic was observed on the fracture surface of the titanium side. Figure  9d shows the crack was initiated at the Al2O3 ceramic and propagated into the brazing seam via the Au/Al2O3 interface during the shear test. The magnified SEM image of Figure 9d is shown in Figure  9e. When the crack propagated into the brazing seam, the joints fractured along the interface of TiAu2 and TiAu4 reaction layers (the interface of zone III and zone IV). The joints brazed at 1115 °C for 1 min fractured in the brittle mode. To further investigate the fracture location, reaction phases on the fracture surface of the titanium side were identified using XRD analysis, as shown in Figure 9c. It was evident that the fracture surface of the titanium side consisted of Au, Al2O3, and TiAu2, which corresponded to the fracture path analyses of Figure 9d,e.  Figure 10 displays the fracture analyses of the joints brazed at different parameters. It was observed that two types of fracture patterns existed after the shear test. In the first fracture pattern, significantly flat fracture surfaces were clearly observed, and the joints fractured along the Au/Al2O3 interface during the shear tests when brazed at 1105 °C for 1 min and 1115 °C for 5 min (Figures 10a  and 9c). XRD analyses of the fracture surface on the titanium side displayed detectable phases, including Au and Al2O3, which in turn supported the above analysis of the first fracture pattern. Meanwhile, as shown in Figure 10b, a second type of fracture pattern was observed when the joints  Figure 10 displays the fracture analyses of the joints brazed at different parameters. It was observed that two types of fracture patterns existed after the shear test. In the first fracture pattern, significantly flat fracture surfaces were clearly observed, and the joints fractured along the Au/Al 2 O 3 interface during the shear tests when brazed at 1105 • C for 1 min and 1115 • C for 5 min (Figures 9c and  10a). XRD analyses of the fracture surface on the titanium side displayed detectable phases, including Au and Al 2 O 3 , which in turn supported the above analysis of the first fracture pattern. Meanwhile, as shown in Figure 10b, a second type of fracture pattern was observed when the joints were brazed at 1115 • C for 3 min, identical with that brazed at 1115 • C for 1 min. In the second type, the fracture started at the Al 2 O 3 ceramic and propagated along the interface of TiAu 2 and TiAu 4 reaction layers, which was confirmed by the existence of Au, Al 2 O 3 , and TiAu 2 in the XRD result.
Variations in shear strength were significant count on the microstructure evolution of the joint. The increase of brazing temperature and holding time can promote the diffusion of active Ti from the titanium substrate and aggregation adjacent to Al 2 O 3 ceramic. When the brazing temperature was lower (or the holding time was shorter), the diffusions of Ti and Au were limited, and the reaction layer of TiO x was extremely thin as the weakest position of the bonding. Therefore, the shear strength of the joints was quite low, and the joint fractured along the Au/Al 2 O 3 interface. With the increase of brazing temperature (or the prolongation of holding time), the TiO x layer thickened, which could improve the metallurgical bonding between brazing alloy and ceramic. Therefore, the shear strength of the joints increased. Fractures occurred at the Au/Al 2 O 3 interface and fragile Ti-Au reaction layers. When the brazing temperature further increased (or holding time was further prolongated), there was a drop in shear strength, which could be attributed to two factors: the over-thickened TiO x layer and the higher stresses resulting from the increased temperature or changed microstructure of Ti-Au IMCs layers in the brazing seam. Based on the above analyses, it can be concluded that a suitable thickness of the TiO x layer adjacent to ceramic had crucial influence on the shear strength of the joints. were brazed at 1115 °C for 3 min, identical with that brazed at 1115 °C for 1 min. In the second type, the fracture started at the Al2O3 ceramic and propagated along the interface of TiAu2 and TiAu4 reaction layers, which was confirmed by the existence of Au, Al2O3, and TiAu2 in the XRD result. Variations in shear strength were significant count on the microstructure evolution of the joint. The increase of brazing temperature and holding time can promote the diffusion of active Ti from the titanium substrate and aggregation adjacent to Al2O3 ceramic. When the brazing temperature was lower (or the holding time was shorter), the diffusions of Ti and Au were limited, and the reaction layer of TiOx was extremely thin as the weakest position of the bonding. Therefore, the shear strength of the joints was quite low, and the joint fractured along the Au/Al2O3 interface. With the increase of brazing temperature (or the prolongation of holding time), the TiOx layer thickened, which could improve the metallurgical bonding between brazing alloy and ceramic. Therefore, the shear strength of the joints increased. Fractures occurred at the Au/Al2O3 interface and fragile Ti-Au reaction layers. When the brazing temperature further increased (or holding time was further prolongated), there was a drop in shear strength, which could be attributed to two factors: the over-thickened TiOx layer and the higher stresses resulting from the increased temperature or changed microstructure of Ti-Au IMCs layers in the brazing seam. Based on the above analyses, it can be concluded that a suitable thickness of the TiOx layer adjacent to ceramic had crucial influence on the shear strength of the joints.  Figure 11 shows nanoindentation test results, which displayed the variation in hardness and elastic modulus of reaction phases for the joint brazed at 1115 °C for 3 min. As shown in Figure 11a, the highest hardness (9.9 GPa) and elastic modulus (165.0 GPa) across the joint was found in the Ti3Au layer, while the Au phase showed the lowest hardness (2.7 GPa) and elastic modulus (115.0 GPa). In order to reveal elastic and plastic behaviors of reaction phases across the joint, typical loads versus depth curves are illustrate in Figure 11b. The deformation process of reaction phases could be divided into elastic-plastic loading and purely elastic unloading. It was apparent that the Au phase possessed the lowest elastic recovery of 14.1%, which recovered 69 nm of the total indentation depth  Figure 11 shows nanoindentation test results, which displayed the variation in hardness and elastic modulus of reaction phases for the joint brazed at 1115 • C for 3 min. As shown in Figure 11a, the highest hardness (9.9 GPa) and elastic modulus (165.0 GPa) across the joint was found in the Ti 3 Au layer, while the Au phase showed the lowest hardness (2.7 GPa) and elastic modulus (115.0 GPa). In order to reveal elastic and plastic behaviors of reaction phases across the joint, typical loads versus depth curves are illustrate in Figure 11b. The deformation process of reaction phases could be divided into elastic-plastic loading and purely elastic unloading. It was apparent that the Au phase possessed the lowest elastic recovery of 14.1%, which recovered 69 nm of the total indentation depth (488 nm). These results showed that the deformation behavior of the Au phase was primarily plastic compared to other phases, which could be beneficial to release residual stress caused by CTE mismatch.
Metals 2019, 9, x FOR PEER REVIEW 10 of 12 compared to other phases, which could be beneficial to release residual stress caused by CTE mismatch.

Conclusions
In this study, a reliable brazing joint of medical titanium and alumina bioceramic was successfully obtained using Au foil. The microstructure, reaction products, and shear strength of titanium/Au/Al2O3 joints were studied. The conclusions are summarized as follows: (1) The typical interfacial microstructure of the titanium/Au/Al2O3 joint was titanium/Ti3Au layer/TiAu layer/TiAu2 layer/TiAu4 layer/Au + granular TiAu4 layer/TiOx phase/Al2O3 ceramic.
(2) Brazing temperature displayed significant effects on the microstructure evolution and mechanical properties of brazed joints. With the increase of brazing temperature, the mutual diffusion of Ti and Au was enhanced, and the thickness of Ti3Au + TiAu + TiAu2 layers adjacent to the titanium substrate increased gradually. Meanwhile, the thickness of the Au layer with granular TiAu4 next to Al2O3 ceramic notably decreased. The TiOX phase, which promoted metallurgical bonding between the brazing alloy and Al2O3 ceramic, could increase as more Ti reacts with Al2O3. The shear strength of the joints increased first and then decreased. When the brazing temperature was 1115 °C, a maximum shear strength was obtained as a result of the TiOX layer with a suitable thickness. Similar effects of holding time on microstructure evolution and mechanical properties were also observed, and the maximum shear strength was obtained for a holding time of 3 min.
(3) Shear tests indicated that the joint brazed at 1115 °C for 3 min exhibited the highest shear strength of 39.2 MPa. Typical fracture analysis displayed that the crack started at the Al2O3 ceramic and propagated along the interface of TiAu2 and TiAu4 reaction layers.

Conclusions
In this study, a reliable brazing joint of medical titanium and alumina bioceramic was successfully obtained using Au foil. The microstructure, reaction products, and shear strength of titanium/Au/Al 2 O 3 joints were studied. The conclusions are summarized as follows: (1) The typical interfacial microstructure of the titanium/Au/Al 2 O 3 joint was titanium/Ti 3 Au layer/TiAu layer/TiAu 2 layer/TiAu 4 layer/Au + granular TiAu 4 layer/TiO x phase/Al 2 O 3 ceramic. (2) Brazing temperature displayed significant effects on the microstructure evolution and mechanical properties of brazed joints. With the increase of brazing temperature, the mutual diffusion of Ti and Au was enhanced, and the thickness of Ti 3 Au + TiAu + TiAu 2 layers adjacent to the titanium substrate increased gradually. Meanwhile, the thickness of the Au layer with granular TiAu 4 next to Al 2 O 3 ceramic notably decreased. The TiO X phase, which promoted metallurgical bonding between the brazing alloy and Al 2 O 3 ceramic, could increase as more Ti reacts with Al 2 O 3 . The shear strength of the joints increased first and then decreased. When the brazing temperature was 1115 • C, a maximum shear strength was obtained as a result of the TiO X layer with a suitable thickness. Similar effects of holding time on microstructure evolution and mechanical properties were also observed, and the maximum shear strength was obtained for a holding time of 3 min.

Conflicts of Interest:
The authors declare no conflict of interest.