**1. Introduction**

Titanium and its alloys have been intensively investigated and applied for biomedical applications because of their excellent biocompatibilities, mechanical properties, and corrosion resistances [1–6]. 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 [7–9]. 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 [10–13]. 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 [9,14–16]. Utilization of metal–ceramic composites for biomedical applications, including implantable pacemakers, retinal implants, and microstimulators, has dramatically increased in recent years [17]. To extend the practical utilization of metal–ceramic composite components, biocompatible metal–ceramic joints are desirable for implantable medical devices [17,18].

Nevertheless, ceramic–metal joints with sufficient mechanical integrity are difficult to achieve because of the large differences in physical and mechanical properties between ceramics and metals such as coefficient of thermal expansion (CTE), chemical composition, and modulus of elasticity (MOE) [10,19]. In the brazing of ceramics to metals, a main problem that needs to be solved is the poor wettability of liquid brazing alloy on ceramics. Active brazing is a promising approach that introduces active elements such as Ti, Zr, Ni, or V into brazing alloys [20–23], which significantly enhances wettability and spreading of liquid metal on ceramics by metallurgical bonding [24].

Ti is a typical active element that promotes wetting and adhesion. The interfacial reactions on the metal/alumina interface have been investigated using various Ti-containing metal alloys such as CuAg-Ti/alumina [21,25,26], AgCu-Ti/alumina [27,28], CuSn-Ti/alumina [29], NiPd-Ti/alumina [30], SnAgCu-Ti/alumina [31], and NiTiZr/alumina [24]. Voytovych et al. identified the existence of M6X-type compounds and titanium oxides (TiOx on a metal/Al2O3 ceramic interface) whose chemical compositions were believed to be greatly dependent on the activity of Ti [25]. Decrease in the activity of Ti in the system leads to the formation of titanium oxides with higher oxidation on the interface, resulting in a higher final contact angle. Similar conclusions were made by Kritsalis et al. after studying an NiPd-Ti/alumina system [30].

It has been widely reported that alumina could be brazed to different metals with Ti-containing filler alloys [10,21,32–37]. The types of titanium oxides that form in an Al2O3/Kovar joint using Ag-Pd/Ti filler are found to be affected by the thickness of the Ti layer, and the joint strengths are influenced by the thicknesses of the reaction layer and residual Ti layer [32]. Xin et al. investigated the reaction products on the Ti film/Al2O3 interface for an Al2O3/Kovar joint and suggested that a competitive reaction mechanism existed in the system. At temperatures lower than 1057 ◦C, Ti reacts with Al from the decomposition of Al2O3, resulting in the formation of Ti3Al. As the reaction proceeds, TiO precipitates out from a Ti solid solution as the O concentration rises above the solubility in Ti and Ti3Al. For test temperatures higher than 1057 ◦C, Ti directly reacts with O from Al2O3 to generate TiO and Ti3Al [10]. Simultaneously, the type of Ti oxide depends on the activity of Ti in the reaction layer, which could be decreased by the interaction between Ti and Ni from Kovar substrate to form Ni3Ti, resulting in a shift of reaction product from TiO to Ti2O3 or Ti3O5 [35]. This is also observed by other investigations [38–40].

The other challenge in brazing ceramics to metals is the thermal stress generated on the metal–ceramic interface resulting from CTE mismatch between them as the joint cools to room temperature. The addition of pure gold, which is biocompatible, is desired because it can release thermal stresses by plastic deformation.

In this study, reliable brazing of Al2O3 ceramic to medical titanium alloy was achieved using pure gold foil. Detailed investigations on the effects of brazing temperature and dwelling time on microstructure evolution and mechanical properties were conducted. Mechanical properties were analyzed from microhardness data for different phases as well as shear strength of the joints.

#### **2. Experimental Materials and Methods**

Commercial polycrystalline Al2O3 ceramic with a purity of 99.5% (Shanghai Unite Technology Co., Ltd., Shanghai, China) was processed into 5 mm × 5 mm × 5 mm cubes. The dimensions of pure α-titanium (Kunshan Bitaite Metal Products Co., Ltd., Kunshan, China) used herein were approximately 20 mm × 10 mm × 5 mm. Figure 1a shows the SEM image of Al2O3 in back-scattered electron (BSE) mode, and Figure 1b displays the metallograph of titanium, indicating the equiaxed structure of the α-Ti alloy. Au foil of purity 99.99% (KYKY Technology Co., Ltd., Beijing, China) with a thickness of 50 μm was used.

**Figure 1.** Microstructures of substrates and schematic diagram of brazing assembly. (**a**) BSE image of Al2O3 ceramic, (**b**) metallographic figure of α-Ti alloy, and (**c**) brazing assembly (mm).

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 <sup>×</sup> 10−<sup>3</sup> 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).
