Influence of microstructure on metal-ceramic bonding in SLM-manufactured titanium alloy crowns and bridges

Materials

The experimental metal powder used is Ti6Al4V (TC4) powder produced by Xi’an Sailong Metal Materials Co., Ltd. The powder has a full, round, and uniform particle size distribution ranging from 19.32 to 52.18 μm, with an average particle size of 22.68 μm and a theoretical density of 4.43 g/cm³. According to the supplier, the powder’s composition (in wt %) was: Al 6.13, V 3.98, Fe 0.036, O 0.081, N 0.02, C 0.008, H 0.0014, balance Ti. The materials used include masking paste (SHOFU Halo MP PASTE OPAQUE) and dental porcelain powder (SHOFU VINTAGE Halo Dentin).

Structural design

Through the simplified design of the surface structure of Calathea zebrina, the structure is made to closely approximate the microstructure of the Calathea zebrina surface within the realm of processability, as shown in the surface structure model and schematic diagram in Figure 1a and b. The wettability of the microstructure system tends to conform to the Wenzel model. Incorporating the design dimensions, the theoretical formula for the surface contact angle is:

Where θ _e represents the intrinsic contact angle of the surface (°); D is the spacing of the microstructure (mm); R is the bottom radius of the microstructure (mm); and H is the height of the microstructure (mm).

Figure 1:

Sample design. (a) Bionic model. (b) Microstructure size diagram. (c) Schematic diagram of the size of the substrate and porcelain.

Two types of samples were prepared: one with a smooth surface and the other with a microstructured surface. The specimens were designed according to the ISO standard ISO 9693:2019, with dimensions shown in Figure 1c. The porcelain area is centered within an 8 mm range, and microstructures are set within this area.

Preparation of samples

The experimental equipment used for sample preparation was the DiMetal-100 multi-material metal 3D printer from Guangzhou Leijia Additive Manufacturing. The optimized process parameters included a laser power of 160 W, a scanning spacing of 0.02 mm, a scanning speed of 500 mm/s, a layer thickness of 0.03 mm, and a spot diameter of 0.08 mm, with an oxygen content of less than 0.1 % during the manufacturing process. Both the control group (smooth surface samples) and the microstructured group samples were ultrasonically cleaned in anhydrous ethanol, acetone, and distilled water for 5 min each. After drying, the surfaces were sandblasted with 50 μm Al₂O₃ particles at a pressure of 5 bar, with the nozzle positioned 50 mm from the samples at an angle of approximately 45°, changing the sandblasting direction every 15 s. After sandblasting, the samples were ultrasonically cleaned in anhydrous ethanol and acetone for 5 min, followed by a 10-min rinse in deionized water before drying. Porcelain application followed the manufacturer’s protocol, with firing parameters as follows: Pre-oxidation (300 °C preheating, 1 s cooling, 50 °C/min heating to 500 °C for 50 s); Opaque Porcelain Layer 1 (500 °C preheating, 10 s cooling, 50 °C/min heating to 760 °C for 50 s); Opaque Porcelain Layer 2 (500 °C preheating, 10 s cooling, 50 °C/min heating to 750 °C for 50 s); Porcelain Body (500 °C preheating, 5 s cooling, 50 °C/min heating to 790 °C for 38 s), the porcelain operation is executed within a vacuum porcelain furnace. After the firing process, the porcelain layer undergoes a polishing procedure to achieve the desired standard geometric configuration.

The standard specimens of each group after porcelain-fused were placed in a universal mechanical testing machine for bending tests. The applied load showed that the cross-loading head was pressed at a constant speed of (1.5 ± 0.5) mm/min until the specimen was broken, and the fracture mutation was recorded. The distance between the support rods is 20 mm, and the diameter of the loading head is 2 mm. The three-point bending strength of the debonding crack initiation strength can be calculated using the following formula:

In the formula: τ _b is the three-point bending strength (MPa); k is the coefficient [24]; F_fail is the measured fracture force. The coefficient k is determined by the thickness d_M of the metal matrix and the Young’s modulus E_M of the metal material. According to the Young’s modulus E_M of the metal substrate, the appropriate curve is selected, and the corresponding k value in the curve is read according to the substrate thickness d_M. The d_M is the thickness of the metal substrate. According to the standard, the d_M is 0.5 mm, the material is TC4, Young’s modulus E_M is about 110 GPa, and the coefficient k is about 4.82.

Experimental design

A single-factor experiment was conducted, with R, D, H, and P as the variables. Samples with varied parameter combinations were fabricated using SLM. Results from the single-factor experiment (Table 1) indicated that the staggered distance had a negligible impact on the actual contact area. Consequently, the staggered distance was found to have no significant effect on the bonding strength of the metal-ceramic interface. To ensure optimal forming quality and microstructure morphology, the subsequent orthogonal experiment focused on a microstructure without any staggered displacement. A three-factor, three-level orthogonal experiment was designed, and the specific levels for each factor are presented in Table 1.

Surface characterization

The microstructures on the sample surfaces were observed using a VHX-600K super-depth microscope (KEYENCE, Japan), and the contact angle values of the surface structures were measured using a video optical contact angle meter (OC2, Germany). The morphology of the metal-ceramic separation surface and the metal-ceramic bonding surface was characterized by SEM and EDS.

The effect of structure size on the bonding strength of metal-ceramic

This study investigated the influence of the microstructure’s bottom surface R, D, H, and P on the bonding strength of metal-ceramic interfaces. Three-point bending tests were conducted on microstructured surfaces featuring varying R, D, H, and P following a uniform porcelain-fusing process to ascertain the optimal outcomes.

Examination of Figure 2 reveals that the surface of the sample after SLM processing exhibits a degree of roughness, indicative of good surface microstructure preparation quality. The density within each group appears uniform, and no significant defects such as cracks or bubbles are observed, which is beneficial for metal-ceramic bonding. Each group of samples demonstrates hydrophilicity, with a CA less than 90°. Generally, a smaller CA correlates with higher bonding strength. As depicted in Figure 2a, the CA initially decreases and then increases with the radius, corresponding to an initial increase and subsequent decrease in metal-ceramic bonding strength. A microstructure with a radius that is too small results in larger gaps between the microstructures, leading to slow droplet spread and ineffective wetting of the microstructure surface. Figure 2b shows that as the spacing increases, the contact angle and metal-ceramic bonding strength follow a pattern of initial decrease and subsequent increase. If the spacing is too small, the molten ceramic powder may not penetrate the narrow gaps, potentially leading to the formation of difficult-to-eliminate bubbles at the base. Conversely, if the spacing is too large, there is a significant reduction in the number of microstructures, which decreases the surface contact area and consequently increases the contact angle. Figure 2c illustrates that when the height of the microstructure is too small, the contact area between the molten porcelain powder and the substrate differs little from that of a smooth surface, with only a slight improvement in wettability. However, if the height of the microstructure is excessively high, the molten porcelain powder struggles to reach the bottom of the gap, leading to bubble formation and difficulty in discharge due to the accumulation of molten porcelain, which in turn reduces the contact area and wettability. Within the experimental range of staggered displacement, the variation in contact angle and metal-ceramic bonding strength is not markedly affected by changes in staggered displacement.

Figure 2:

Contact angles and bonding strength curves of different structure sizes (a) R, (b) D, (c) H and (d) P.

The orthogonal experiment method was used to design the orthogonal experiment of three factors and three levels. According to the results of the single-factor experiment, the level variables of each factor were set. Table 2 records the contact angle of the orthogonal experiment and the results of the three-point bending test.

Table 3 presents the results of a multi-factor analysis of variance examining the influence of microstructure R,D and H on metal-ceramic interface bonding strength. The model’s coefficient of determination (R²) was 0.984, indicating that spacing and height significantly influence the bonding strength, with a significance level of 0.05 (p<0.05). The range analysis revealed that the impact of these factors on the bonding strength was in the order of H>D>R. Specifically, under the specific parametric conditions where the microstructural R is precisely 0.17 mm, the interstitial D is 0.80 mm, and the geometric H is 0.20 mm, in conjunction with the absence of any staggered displacement, the CA of the microstructure surface was determined to be 10.1°. Concurrently, the metal-ceramic bond strength achieved a maximum value of (51.06 ± 1.22) MPa. These optimal conditions not only enhanced the surface quality of the fabricated components but also yielded the highest metal-ceramic bond strength, as graphically depicted in Figure 3.

Figure 3:

Surface SEM image and contact angle of optimal microstructure parameters.

Analysis of metal-porcelain separation surface

Figure 4a illustrates that the fracture surface of the microstructure-free group exhibited flake-like failure, with a substantial number of porcelain pieces completely detached, leaving minimal porcelain bonding residue. A substantial area of the titanium alloy matrix was exposed, indicating that the majority of the porcelain had peeled off. In contrast, Figure 4b shows that within the micro-pits created by sandblasting, there was a limited presence of porcelain residue, confined to spot-like adhesive areas, while the remainder displayed flake detachment. Figure 4c demonstrates that on the microstructured surface, a significant amount of porcelain is embedded within the microstructure gaps, resulting in a high proportion of porcelain residue. Owing to the microstructure’s unique wettability, numerous molten porcelain powder particles adhere to the gaps of each microstructure unit during the baking process. Facilitated by the enhanced wettability and an extended contact area, the porcelain layer can tightly adhere to the microstructure surface, forming mechanical interlocks and improving the mechanical bonding force at the bonding interface, thereby enhancing the metal-porcelain bonding strength. The microstructured substrate can markedly improve the metal-ceramic bonding strength, with the failure mode of the metal-ceramic interface primarily being adhesive failure.

Figure 4:

SEM images of the failure surface of metal-ceramic. (a) Smooth surface. (b) Porcelain residue. (c) Optimal microstructure surface. (d) Local amplification at the microstructure.

EDS analysis was performed on the fracture surface to ascertain the proportion of ceramic residue and the elemental distribution, as depicted in Figure 5a and b. The common failure modes in the three-point bending test of metal-ceramic structures are categorized into three types: (1) Cohesive failure: where porcelain cracks occur within the porcelain layer itself; (2) Adhesive failure: where the interface between the metal and porcelain layers is compromised; (3) Mixed-mode failure: involving a combination of the above two types of porcelain fractures [25], 26]. Analysis reveals that the low bonding strength between metal and porcelain on the microstructure-free surface leads to extensive flaking of the porcelain layer, with minimal detectable ceramic powder elements, suggesting that the fracture occurs in the relatively weaker oxide layer, with ceramic residue below 20 %. This indicates adhesive failure. In contrast, samples with microstructures exhibit mixed-mode failure. Optimization of the microstructure results in a significant increase in Si residue on the metal-ceramic fracture surface, reaching approximately 72.23 %, with the fracture occurring through both the oxide layer and the porcelain layer. This represents a 55 % increase compared to the non-microstructured surface, demonstrating a superior mechanical interlock and enhanced bonding strength between the metal and porcelain.

Figure 5:

EDS scanning results of separation surface (a) sandblasting surface and (b) optimal microstructure surface. SEM images of metal-ceramic bonding section (c) no microstructure group and (d) optimal microstructure group.

Analysis of metal-ceramic bonding cross section

Examination of Figure 5c reveals that the average thickness of the oxide layer at the metal-ceramic interface on the microstructure-free surface is excessively thick, surpassing 5 μm. The presence of a dense and robustly adhered oxide layer on the metal surface is fundamental to enhancing the bonding strength between metal and ceramic. However, the presence of micropores at the interface and a multitude of unmelted ceramic powder particles during the sintering process can lead to an increased oxidation potential at the interface, causing a marked increase in oxide layer thickness. This unceasing oxidation of the substrate metal, due to the insufficient barrier properties of the oxide layer, results in a continuously growing oxide thickness that adversely affects the wettability of the metal-ceramic bonding interface. The oxide layer adhering to the metal surface is characterized by an intermittent and loosely packed morphology, which diminishes the chemical bonding efficacy at the interface. In contrast, Figure 5d demonstrates that the microstructured surface with optimal parameters exhibits excellent wettability, limiting the bonding interface’s exposure to excessive oxygen. Consequently, the oxide layer thickness at the metal-ceramic interface is comparatively slender, measuring 0.5–0.7 μm, with a uniform distribution. The oxide layer is intimately integrated between the metal and the ceramic layer. Partial melting of the metal surface allows for a tight interlocking with the underlying ceramic layer, creating a mechanical locking effect akin to that desired in bonding. The bonding interface is devoid of defects such as micropores, presenting a smooth and even surface topography.