Abstract
Objectives
To compare the bond strength of a ceramic veneer material to two different ceramic core materials using shear strength testing and finite element analysis (FEA).
Methods
15 blocks of ceramic cores were made of each of IPS e.max ZirCAD (Ivoclar-Vivadent, Schaan, Liechtenstein) (ZirCAD) and lithium disilicate IPS e.max CAD (LS). These were veneered using IPS e.max Ceram (Ceram) according to the manufacturers’ instructions. Samples were then mounted, placed inside a shear testing jig and loaded at their core/veneer interfaces with a universal testing machine. Fractured specimens were then examined fractographically at 45×. An independent t -test and a Mann–Whitney U test were used to detect differences in bond strengths and the patterns of failure between the two groups, respectively ( p > 0.05). Two-dimensional plane stress FE-models were constructed and subjected to shear loading simulating the experimental conditions.
Results
Mean shear stresses (MPa) were 28.8 (9.5) for ZirCAD/Ceram and 29.1 (8.3) for LS/Ceram. Differences were not statistically significant. Fracture patterns were significantly different between the 2 groups as all ZirCAD/Ceram samples broke adhesively at their interface while LS/Ceram samples broke cohesively in the veneer or the core or had mixed adhesive/cohesive failure. Stress distributions in the FEA models were also different, corresponding to the variable fracture patterns in the 2 groups.
Significance
Although shear stresses were similar, fractographic analysis and finite element modeling suggested better bonding between the veneering ceramic and the glass–ceramic than to the zirconia cores. Improved bonding techniques are necessary to prevent clinical delamination of veneered zirconia restorations.
1
Introduction
Due to their inherent brittleness, ceramic indirect restorations usually consist of a core of unesthetic but strong ceramic material veneered with weaker but more translucent ceramic. Fracture propagates through the weakest component in these structures, which is usually the veneer or the core/veneer interface, resulting in delamination of the veneer or extending through the core . The presence of complex intra-oral tensile stress distributions in addition to inherent structural flaws might eventually cause failure of these restorations at low failure loads. Therefore, bond strength studies of ceramic core/veneer combinations can help predict the chances for survival or chipping of such materials in clinical situations and may assist clinicians in the selection of the most durable combinations .
Despite the abundance of the studies investigating bond strength of different interfaces, a universally accepted bond strength protocol has not yet been found in the dental field . Furthermore, although the search for a standardized and ideal bond strength test was the subject of discussion in the Academy of Dental Materials meeting in 2009 in Portland, Oregon, no single test was agreed upon but among the different methods used, shear testing was the most popular . Under shear stress, the layers of a particular material may slide over each other and in shear strength tests, two materials are joined using an adhesive layer and force is applied at the interface till separation of the two layers occurs . Alternative tests include a 4-point flexural test , extrusion shear strength test , punching shear test , or tests using more clinically relevant scenarios as in pull-off tests of ceramic crowns cemented to natural extracted human teeth . Although the tensile test has been commonly used , it can result in non-uniform stress distributions and it needs a complicated sample preparation . If not properly controlled, the specimens in tensile testing undergo torque stress reducing the bond strength. The micro-tensile test, on the other hand, uses very small sample dimensions and is claimed to generate a uniform stress distribution . However, the micro-tensile test is generally difficult to conduct and is easily affected by many variables. Although micro-shear is easier and simpler to perform than the micro-tensile test , it was not proven superior to the more conventional shear test .
Shear testing is relatively easy to perform and is useful in predicting the performance of materials . However, as the real situation is more complex, in vitro results should be interpreted with caution . In any case, the quality of the bond should not be based solely on nominal values of bond strength , but analysis of the failure mode and fractographic analyses of the debonded surfaces can provide valuable information regarding the performance of the materials under forces experienced clinically . A model simulation using FEA can also help analyzing stress distributions in clinical situations . Furthermore, standard testing parameters clarifying specimen design and test configuration should be implemented in shear testing protocols in order to compare the results of studies performed by different researchers .
Several studies have investigated the bond strength of ceramic veneers to ceramic cores. Dundar et al. investigated shear and microtensile bond strengths of four different core/veneer combinations and found significant differences among the various material combinations . Aboushelib et al. evaluated the core/veneer bond strength of different bilayered ceramics using a microtensile test, SEM fractography and FEA. The results showed that bonding of IPS Empress 2 to its veneer was significantly stronger than that of zirconia to its veneer .
Aboushelib et al. found improved bonding between pressable veneers with zirconia cores compared to the conventionally layered veneers . A double veneering technique was later found to improve the esthetic quality of pressable veneers, while simultaneously maintaining their superior bonding to zirconia . More recently, Beuer et al. used an innovative technique for veneering zirconia by fabricating CAD/CAM lithium disilicate veneers that are sintered to zirconia cores and joined by low-fusing ceramic material . Although sandblasting of zirconia cores was suggested , neither sandblasting, heat treatment, liner application nor silica-coating procedures enhanced bonding to zirconia. This suggests that these ceramic composites can only be improved if a stronger veneer is used .
Bond strength and quality between the veneer and the core ceramic are key factors in the success of bilayered restorations. However, this field still requires further investigation. Only a few studies are available evaluating newly released veneer/core combinations. Therefore, the aim of this study was to investigate the shear bond strength of a ceramic veneer material to two different ceramic core materials using both experimental and analytical methods.
The objectives were:
- 1.
To compare the shear bond strength of two ceramic core/veneer interfaces.
- 2.
To analyze the mode of failure of the two groups.
- 3.
To use finite element models to analyze stress distributions under similar testing conditions.
The null hypotheses were:
- 1.
There are no differences in either bond strength or the patterns of failure between the two groups investigated.
- 2.
There are no differences in the patterns of stress distribution in the finite element models of these material combinations.
2
Materials and methods
2.1
Specimen preparation
Two material combinations were tested in this study ( n = 15) (ZirCAD/Ceram and LS/Ceram). Each specimen was composed of a rectangular base of ceramic material (12.5 mm × 11.0 mm × 0.9 mm) to which a block of ceramic veneer (2.5 mm × 11.0 mm × 0.9 mm) was bonded at one end. Table 1 shows the materials used and their composition. Rectangular-shaped samples were fabricated in this study since the ceramic core materials were supplied by the manufacturers as blocks and so rectangular sections were easily cut and prepared. Anusavice et al. showed that the rectangular parallel shear test, using samples with similar geometries to those used in the current study, had the lowest stress concentration effect and the least possibility to cause fracture due to tensile stress concentrations, compared to other testing geometries . In this study, dimensions of the samples were standardized and were relatively small as smaller bonded areas are expected to have fewer defects than bigger areas and hence higher bond strength and a more uniform stress distribution .
| Material a | Basic composition | CTE (10 −6 K −1 ) 100–500 °C |
|---|---|---|
| IPS e.max ZirCAD ® (ZirCAD) | Glass-free densely dispersed 98 vol% tetragonal fine grain ZrO 2 (<0.5 μm) metastabilized with 2% of Y 2 O 3 | 10.8 |
| IPS e.max CAD ® (LS) | Glass–ceramic with 70 vol% lithium disilicate crystals (Li 2 Si 2 O 5 ) | 10.45 ± 0.25 |
| IPS e.max Ceram ® (Ceram) | Glass–ceramic veneer with 8.5 vol% fluorapatite crystals (Ca 5 (PO 4 )3F) | 9.5 ± 0.25 |
ZirCAD bases were prepared by cutting sections of ZirCAD “green body blocks” of dimensions of (15 mm × 14 mm × 1.1 mm) using a slow-cutting diamond saw with water irrigation. The rectangular specimens were sintered in a Sintramat furnace (Ivoclar-Vivadent, Schaan, Liechtenstein) resulting in 20% shrinkage of their original volume. LS bases were prepared by cutting sections of LS blocks. The rectangular specimens were then fully crystallized in a Programat EP 5000 furnace (Ivoclar-Vivadent, Schaan, Liechtenstein). All bases had the final average dimensions of (12.5 mm × 11 mm × 0.9 mm).
Following sintering of the ZirCAD cores, a layer of ZirLiner (Ivoclar-Vivadent, Schaan, Liechtenstein) was applied and fired according to the manufacturers’ instructions to enhance bonding between zirconia cores and the veneer. This was followed by veneering with Ceram material starting from a wash, followed by dentin and finally incisal ceramic, as per a typical clinical situation. The veneer was applied in a layering technique to a thickness of 2.5 mm as shown in Fig. 1 . All excess veneering material was trimmed and samples were polished so that the veneer/core interface was clearly visible. Veneering LS cores was similar except that Ceram was applied directly on the core without ZirLiner. Although the dimensions of the samples were standardized, slight variations were unavoidable. Therefore, the dimensions of the bonded surface area for each sample were measured. Samples were then stored in water at 37 °C for 24 h.
Samples were then prepared for shear loading, with a Zwick/Roell Z020 universal testing machine (Zwick GmbH & Co. KG, Ulm, Germany), by mounting them in brass cylinders with the ceramic base of each sample fully-embedded in stone, to stabilize the samples in that exact position, while the veneer projected out of the cylinder. Every effort was made to have the core/veneer interface at the same level with the rim of the brass cylinder. Stone was chosen as it is easier to remove later from the cylinders and from around the samples which were retrieved after testing for analysis of fracture patterns under the microscope. Each sample was placed in a shear-test jig ( Fig. 1 ). The jig secured the samples in place while the loading edge contacts the veneer at a distance less than 0.5 mm from the core/veneer interface. A crosshead speed of 0.2 mm/min was used until failure occurred. Although a crosshead speed of 0.5–0.75 was recommended by the ISO Standards and used in many studies , the effect of crosshead speed was found minimal and in this study, a speed of 0.2 mm/min was selected, as it was thought to produce more accurate data, especially since the samples were small and also based on previous research by Kukiattrakoon and Thammasitboon .
Maximum load to failure was recorded for each sample. The shear stress ( τ ) was obtained by dividing the force to failure ( F ) by the surface area ( A ) of each sample:
2.2
Fracture surface analysis
Fractured specimens were retrieved from the testing assembly and examined using a Meiji EMZ-TR optical microscope (Meiji Techno Co., Ltd., Tokyo, Japan) with a maximum magnification of 45×. Failure was classified into four categories:
- 1.
Cohesive failure in the veneer.
- 2.
Mixed fracture starting just near the interface, crossing it and continuing in the core.
- 3.
Adhesive failure at the core/veneer interface.
- 4.
Cohesive failure in the core.
An independent t -test and a Mann–Whitney U test were used to detect statistically significant differences in the bond strength values and the pattern of failure between the 2 groups, respectively.
2.3
Finite element analysis (FEA)
A two-dimensional plane stress FE-model ( Fig. 2 ) was created to simulate specimens used in the experimental shear test, since the specimens were thin enough such that the variations of displacement with respect to the Z -axis were negligible and stresses in this direction were zero (small order of magnitude). MSC PATRAN (v. r2 2005, MSC Software, CA, USA) software was used to perform the pre- and post-processing. MSC Marc (v. r2 2005, MSC Software, CA, USA) was used as a solver.
The geometry of the models was characterized by drawing points and curves with the FE software, which generated surfaces with the same dimensions of the specimens used in the experimental shear test. These surfaces were divided into a finite number of elements to create the mesh. Four-node arbitrary quadrilateral elements ( Paver ) were used to build this mesh, constituted of 4307 elements and 4440 nodes. A higher density of meshing was used at the core/veneer interface, the adjacent areas of the core and the veneer in addition to the upper surface of the veneer close to the points of load application as well as the lower surface. The model was constrained in the X and Y directions along the rows of nodes at the outer edges of the ceramic base part. The part of the model representing the veneer was left free to move under loading, resembling the situation studied experimentally.
The materials were assumed to be elastic, isotropic, homogenous and linear-elastic. Table 2 shows the elastic moduli ( E ) and Poisson’s ratios ( υ ) for the materials used in this study, as provided by the manufacturers and previous studies . Two models were constructed representing the two groups. A three-node concentrated loading condition was simulated at a distance of 0.15 mm from the core/veneer interface. An arbitrary 150 N load was applied (50 N per each node), based on the mean value recorded by pilot studies. Since this study was focused on the stress distributions, the absolute stress values were not considered. Therefore, the selected load only affected the stress magnitude, not the stress distribution. The first interfacial node in the models was not loaded and load was applied at the three successive nodes on the top surface of the veneering ceramic component of the model. This was done in an attempt to mimic the situation most probably created in the experiment as there were inevitably limits to the tolerance with which the samples could be positioned with their interfaces exactly at the level of the rims of the brass cylinders. Furthermore, the loading component was around 0.4 mm thick and the distance between each node in the model was 0.15 mm. Consequently, load was applied on three nodes rather than one node.
| Material | E (GPa) | υ |
|---|---|---|
| IPS e.max ZirCAD ® (ZirCAD) | 220 | 0.30 |
| IPS e.max CAD ® (LS) | 95 | 0.23 |
| IPS e.max Ceram ® (Ceram) | 68 | 0.24 |
Both shear and tensile stress distributions were analyzed in each model. The ratios shear/tensile stress were calculated and used in further analysis for the prediction of failure. The greater shear stress presents at the highest stress region, the higher the tendency for adhesive failures to occur, since the absence of tensile stress at the highest stress region means that the bonding at the interface was much lower than the cohesive bonding.
2
Materials and methods
2.1
Specimen preparation
Two material combinations were tested in this study ( n = 15) (ZirCAD/Ceram and LS/Ceram). Each specimen was composed of a rectangular base of ceramic material (12.5 mm × 11.0 mm × 0.9 mm) to which a block of ceramic veneer (2.5 mm × 11.0 mm × 0.9 mm) was bonded at one end. Table 1 shows the materials used and their composition. Rectangular-shaped samples were fabricated in this study since the ceramic core materials were supplied by the manufacturers as blocks and so rectangular sections were easily cut and prepared. Anusavice et al. showed that the rectangular parallel shear test, using samples with similar geometries to those used in the current study, had the lowest stress concentration effect and the least possibility to cause fracture due to tensile stress concentrations, compared to other testing geometries . In this study, dimensions of the samples were standardized and were relatively small as smaller bonded areas are expected to have fewer defects than bigger areas and hence higher bond strength and a more uniform stress distribution .
