INTRODUCTION

Metal-free restorations are a suitable clinical option for indirect procedures because of their biocompatibility, excellent mechanical properties, and optimal esthetics.¹ In view of these properties, lithium disilicate glass-ceramic can be considered the material of choice for many clinical situations. To achieve a proper bonding between inorganic restorative materials and organic tooth tissues, coupling agents capable of linking both surfaces must be applied to them.^2-4 Moreover, the use of resin cement to bond ceramic restorations to teeth may improve the mechanical performance at the tooth-ceramic junction.^5,6 To ensure long-lasting tooth-restoration bonding, all these issues must be considered.⁷

Glass-ceramic surface treatment must be performed in order to enhance resin cement/glass-ceramic bonding. Two main approaches are recommended in the literature: mechanical and chemical.² The mechanical strategy consists of etching the ceramic surface using hydrofluoric acid (HF), which produces a selective glass-content removal, exposes the crystalline structure, raises surface energy, and facilitates mechanical interlocking of the resin cement.^4,5,8,9 Otherwise, silane couplers provide chemical adhesion between resin cements and silica-containing ceramic substrates.⁵ Silane molecules employed in dentistry contain two functional groups: one reacting with polymerizable methacrylates, the other reactive toward silica in glassy structures. The alkoxy groups of this molecule must be activated by a hydrolyzation process (SiOR→SiOH), being suitable then to undergo a condensation reaction when in contact with a ceramic surface, in which water is released as a byproduct.² The methacrylate group reacts with the polymerizable side of the resin cement² to achieve a three-dimensional, cross-linked network between the ceramic and resin cement.^10-12 This chemical process, complemented with mechanical interlocking, is currently the most accepted procedure for enhancing resin cement/glass-ceramic bonding.^2,4,5

Following a simpler approach for dental bonding, manufacturers have also added silane to dental adhesives, specifically to a category of materials known as “universal adhesives.” These normally contain phosphate acid monomers, such as 10-methacryloyloxydecyl dihydrogen phosphate (MDP), which can chemically bond with hydroxyapatite (making it possible to eliminate dentin etching with phosphoric acid), metallic ions present in some ceramics (mainly polycrystalline ceramics), and methacrylate groups of resin cements.^13-15 The addition of silane makes a chemical interaction with glass-containing ceramics possible, amplifying the range of substrates in which these adhesives will act as bonding promoters for resin-based materials.¹⁶ Derived from this, manufacturers claim that silane-containing universal adhesives can be used as conventional silane primers. However, some studies have reported that universal adhesives produced lower ceramic-cement bond strength than conventional silane primers or separate use of silane and adhesive,^17-19 probably because of some kind of silane inactivation inside universal adhesives of low pH or other reason.^19-20

This broad range of surface treatments, showing different effectiveness, may not only affect glass-ceramic/resin cement adhesion, but its mechanical properties as well. Related to this, the influence of some surface treatment protocols on a material's mechanical performance has also been evaluated. Previous studies have shown an increase in ceramic flexural strength when applying an adhesive, unfilled resin coat,^21-24 or resin cement¹² after a silane coupler. Conversely, another investigation found that the application of conventional silane alone (not resin-cement coated), exerted no effect on ceramic biaxial flexural strength, being more involved with the ceramic's surface texture and unfilled resin application.²⁵ However, the effect of different types of silane primers (probably showing dissimilar bonding-promoting effectiveness between them) on the flexural strength of ceramic/cement systems is still uncertain.

Thus, the way luting procedures are managed may affect a restoration's bonding and mechanical performance, which are important parameters in understanding clinical behavior.²⁶ Currently, many options are available to perform glass-ceramic silanization, and manufacturers recommend that they be employed indistinctly (whether silane is mixed with other components or not). To the authors' knowledge, no previous study has evaluated the effect of silane-containing universal adhesives on biaxial flexural strength of a glass-ceramic/resin cement system or the bonding stability they provide to the adhesive interface when submitted to loading forces.

Here we determine whether the use of a silane-containing universal adhesive such as silane primer affects the biaxial flexural strength (BFS) and bonded interface integrity of glass-ceramic/resin cement systems after loading. The null hypotheses tested were that silane-containing universal adhesive does not influence 1) a glass-ceramic/resin cement system's biaxial flexural strength or 2) a glass-ceramic/resin cement system's adhesive interface stability after loading.

METHODS AND MATERIALS

Specimen Fabrication and Group Division

Lithium disilicate CAD/CAM blocks (IPS e.max lithium disilicate CAD/CAM, A2 color, Ivoclar, Vivadent, Schaan, Liechtenstein) were milled on an E4D Dentist System using a cylindrical custom-milled file measuring 6.5 ± 0.1 mm in diameter. Each cylinder was then cut with a diamond saw under water irrigation to obtain discs of 0.5 ± 0.1 mm thickness until completing 120 disc-shaped specimens. Sample measurements were then matched to the appropriate diameter (6.5±0.1 mm) using a digital caliper to fit the biaxial flexure jig. Specimen and flexure jig device dimensions were chosen to simulate the approximate size of a ceramic veneer, as employed in a previous work.²⁷ Discs were fired unglazed according to the manufacturer's instructions and then ground using #1000 and #2000 grit silicon carbide grinding paper etched with 5% hydrofluoric acid (HF) (Power C Etching, BM4, Palhoça, SC, Brazil) for 20 seconds, water-cleaned for 60 seconds, and ultrasonically cleaned for 5 minutes.

Four groups were formed employing different silanization protocols, treating specimens as follows: 1) RCP (conventional silane): One coat of RelyX Ceramic Primer (3M ESPE, St Paul, MN, USA), a one-bottle conventional silane primer, was actively applied onto the ceramic surface for 60 seconds, followed by thorough drying (20 seconds) using oil-free air until complete solvent evaporation; 2) RCP+SB (conventional silane plus separate adhesive application): one coat of RCP was also applied onto the ceramic surface for 60 seconds and then thoroughly dried (20 seconds) using oil-free air until complete solvent evaporation. Afterward, an adhesive system (Adper Single Bond Plus, 3M ESPE) was applied in one coat for 15 seconds and air-dried for 5 seconds to evaporate the solvent; 3) SBU: Scotchbond Universal (3M ESPE) was applied onto the ceramic surface in one coat for 20 seconds and air-dried for 5 seconds; 4) C: No silane was used and only the previously described HF etching procedure was performed on this group.

After being silanized, all treated surfaces were resin-cement coated (RelyX Ultimate, 3M ESPE). To do so, treated specimens were fixed by the untreated surface on a thick glass plate with the aid of utility wax. One layer of resin cement was placed on the treated surface with the aid of an auto-mixing tip provided by the manufacturer and a micro brush. A polyester strip and a 0.5-mm glass slide were placed on top of the resin cement and pressed using standardized weight devices (200 g) at each side of the glass slide, and at the same time, attached to a digital caliper to control specimen thickness. The resin cement layer was polymerized for 40 seconds (Elipar, S10, 3M ESPE; 800mW/cm² as determined using an Ophir Laser measurement potentiometer from Ophir Optronics Ltd., Jerusalem, Israel and taking into account light tip circular area). In the case of groups treated with adhesive systems (RCP+SB and SBU), adhesive and resin cement layers were polymerized simultaneously. Specimen thicknesses were measured again using a digital caliper, and the readings were recorded to calculate each layer's thickness for each specimen (all thicknesses were confirmed using the fractured fragments with the aid of a scanning electron microscope (SEM). Calibration of specimens to fit the flexure jig was performed using a 2000-grit silicon carbide grinding paper whenever necessary. Materials used are described in Table 1.

Table 1 Materials Used and Their Application Steps

View Fullscreen

Biaxial Flexural Strength Test

To measure BFS, the piston-on-ring method was used, employing a customized flexure jig.²⁷ A schematic representation of the biaxial flexural strength test used in this study is presented in Figure 1. After 24 hour storage in 100% relative humidity at 37°C, all specimens were loosely fitted (resin cement layer facing down) on a support ring (5 mm internal diameter) through a circular aperture (7 mm diameter) of a cylindrical stainless steel jig. Slight specimen flatness imperfections were offset by using a thin piece of rubber film along with a wet piece of filter paper.^28,29 The assembly was positioned on a universal testing machine working at 1.27 mm/min (Instron 4411, Instron Corp, Canton, MA, USA), and a vertical load was applied on the middle of the specimen until failure by a circle-shaped flat piston. The process was monitored and the load recorded at the point of failure, which was used to calculate the BFS/σ_θ according to the bilayer disk approach (considering ceramic and cement as different layers). This method was proposed by Hsueh and others,^29-31 using the analytical model of bilayered disks tested on piston-on-ring device, described by equations 1 to 4³⁰:

(for t₁ ≤ z ≤ t₁+ t₂ and r = c),

where P is the load (N) at fracture, E₂ is the individual Young modulus of layer 2: ceramic (102.7 GPa³²), z is the axial position of the desired point of calculation on the vertical axis (in this case the axial position used was z = t₁ [ceramic/cement interface]), a is the support ring radius (2.5 mm), c is the radius of the indenter of the piston (0.8 mm), and R is the specimen radius (3.25 mm) (Figure 1). The variable ν is given by

which is an average considering Poisson ratios from each material (0.215 for ceramic [v₂][specifically for E.max CAD³²], and 0.27 for resin cement [v₁]³³) and each layers' thickness (t₁: individual resin cement layer thickness; t₂: individual ceramic layer thickness). Each specimen's thickness was measured individually. Variable z* represents the position of the neutral plane and is given by

where E₁ is the Young's modulus of layer 1: resin cement (10 GPa²⁸) and variables t₁, t₂, v₁, v₂ and E₂ are the same as used in equations 4 and 5. D* is the flexural rigidity, described by

View Figure

• Download as Powerpoint
• Download Figure

All fragments were collected, identified by specimen, and the number of fragments obtained from each sample was recorded.

RESULTS

Statistical analysis showed that the data were normally distributed (Anderson-Darling test [p=0.325]); also, homoscedasticity was proved (Bartlett test [p=0.426]), both at a preset alpha of 0.05, indicating allowable use of parametric methods for data analysis. One-way ANOVA statistical analysis revealed that the factor “silane treatment” significantly influenced BFS (p<0.0001). RCP+SB showed the highest BFS mean, showing no statistical difference from the one obtained by RCP (Table 2). SBU presented a lower BFS mean value than RCP+SB and RCP, but not different from C (Table 2).

Table 2 Biaxial Flexural Strength Values (BFS) With Standard Deviation (SD) and Weibull Modulus (m) With Confidence Intervals (CI) From All Experimental Groups

View Fullscreen

No statistical differences were found in m, as all confidence intervals overlapped at least at one point (Table 2). Although not statistically significant, the highest m/graph slope was obtained by SBU and the lowest m/graph slope by the C group (Table 2 and Figure 2).

View Figure

• Download as Powerpoint
• Download Figure

Representative images from fractured specimen analysis are summarized in Figures 3 to 6. Fractured fragment analysis revealed ceramic-cement separation for groups SBU and C (Figures 5 and 6). In the particular case of SBU (Figure 5), the failure line was located mostly between the adhesive layer and the ceramic material. Conversely, groups RCP and RCP+SB showed a uniform interlocking area (resin cement/ceramic [RCP] and resin cement/adhesive/ceramic [RCP+SB/RC]), in which no gap or interruption of continuity were noted (Figures 3 and 4). Additionally, some voids were noticeable within the resin cement layer of all groups.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

DISCUSSION

This study shows that silanization using a silane-containing universal adhesive produced lower BFS values on ceramic/cement systems than conventional silane and separate application of silane and adhesive. Consequently, the first null hypothesis must be rejected. Because the use of silane-containing universal adhesive as ceramic primer negatively affected the integrity of ceramic/cement adhesive interface during loading, null hypothesis 2 was also rejected. Thus, it can be said that regarding a glass-ceramic/resin cement assembly's mechanical properties, the way resin cement is bonded to the ceramic material may be a very relevant aspect. Additionally, the role of an efficient silane coupling agent appears to be fundamental in maintaining the integrity of the materials composing the ceramic indirect restoration system.

In order to analyze the mechanical properties of a glass-ceramic/resin cement system influenced by different types of silane primers, the biaxial flexural strength was evaluated using a piston-on-ring biaxial flexure strength test. The biaxial flexure test is more reliable than uniaxial tests, as it applies the stress on a concentric point of the specimen, resulting in a more uniform analysis of material strength.^24,34 The traditional approach for this kind of bending test is not useful to calculate the BFS of multilayered specimens composed of more than one dissimilar material, as it fails to consider the Poisson ratio and individual thickness of each material.^29-31 The analytical solution proposed by Hsueh and others^28-30 includes this possibility, and it is also efficient in calculating biaxial stresses through ceramic/cement bilayered specimens.²⁸ In the particular case of ceramic specimens treated with an adhesive system layer before resin cement coating (RCP+SB and SBU), although these specimens are composed of three different materials, they were treated as bilayered specimens for calculation purposes, as adhesive layer thickness in this study was recorded to be under 15 μm in all specimens. Such a small thickness would not influence the calculation outcomes.²⁸ Furthermore, as adhesive and resin cement layers are both composed of resin-based materials and were polymerized together, they were observed to be well integrated; sometimes it was even difficult to differentiate between the two. Thus, being equally valid to apply the bilayer approach for those situations,^24,28 adhesive layers were considered as part of resin cement layers in such cases.

The application of conventional silane (alone or adhesive coated) improved BFS, indicating that silane efficiency in enhancing ceramic/cement bonding plays an important role in a ceramic/cement assembly's mechanical properties. This effect might be explained by the fact that silane bonds chemically with glass-ceramic and resin-based materials, maintaining the integrity of the system. This can be confirmed on SEM images, where a full continuity of the ceramic-cement or ceramic-adhesive-cement interfaces is observed (Figures 3 and 4), suggesting this chemical union was successful. This uniform interlocking area between ceramic and luting materials was previously reported to strengthen glass-ceramics.^21,25 Addison and colleagues suggest that, as an explanation to this phenomenon, when a crack is filled by the resin-based material and the whole system is submitted to load, the Poisson effect is expected to occur, producing a slight contraction at the bottom of the crack (due to its geometry), raising the elastic modulus of the resin-based material in this area and equalizing its behavior to that of the ceramic material.^21,25 This cumulative effect within adjacent cracks may strengthen the resin-based material, maintain ceramic/resin unity, and as a consequence, increase the flexural strength of the whole system.^21,25 On the other hand, applying an adhesive coat after the silane primer may present better wetting and interpenetrating capabilities than would the resin cement due to the adhesive's lower filler content and viscosity, resulting in a better intimacy with the ceramic material.^23,35 Thus, in light of our results, adhesive coating of previously silanized glass-ceramics has a positive effect on the ceramic/cement system's mechanical properties.

Silane-containing universal adhesives are also recommended by manufacturers to be used as silane primers. In the present study, this kind of material demonstrated no positive effect on the biaxial flexural strength, as it showed no statistical difference compared with the negative control group (Table 2). This may be due to some kind of inefficient bonding property demonstrated by the silane contained in these adhesives, incapable of maintaining the integrity of the ceramic-adhesive-cement assembly, as shown in Figure 5. This lack of unity between “bonded” materials may lead to a faster propagation of microcracks, consequently weakening the whole specimen.

In this study, this pattern was observed in the SBU group and in the negative control group in which silane was not applied (Figures 5 and 6). Previous studies have demonstrated that conventional silane (with or without separate applications of adhesive) performs better than silane-containing universal adhesives as a ceramic-cement bonding promoter.^17-19 Universal adhesives contain many ingredients other than silane, resulting in fewer silane molecules per area in contact with the ceramic surface,³⁶ in contrast to the silane-only containing primer. Intimate contact between the silane and ceramic surface is crucial, as one silane coat contains three oligomer layers,³⁷ just the first being capable of forming chemical bonds; the outermost layers may be detrimental.³⁸ Also, elimination of solvents and other byproducts formed during the silane condensation reaction may be hindered through development of a dense polymer network,³⁹ needing more time to evaporate solvent in universal adhesives as demonstrated in a previous work.⁴⁰ Moreover, a more acidic environment of universal adhesives (SBU, pH 2.7; RCP, pH 4.6)² may lead to continuous hydrolyzation and reaction of its silane molecules upon storage, and consequently being inactive to some degree before being used, as proven by Yoshihara and others.²⁰

Additionally, the type of silane contained in SBU is not specified; it is possible that the silane compound used in those adhesives is not as effective as the well-known methacryloxypropyltrimethoxysilane. All these issues may explain why universal adhesives failed to improve the ceramic/cement system's BFS and maintain its integrity with resin cement coating. However (though not statistically significant), the SBU attained the highest m (16.6) among all groups (Figure 2, Table 2), being very distant from the C group in this regard. This may be explained by the fact that, despite the similar behavior that they (SBU and C) showed regarding BFS and adhesive interface integrity (Table 2, Figures 5 and 6), SBU as an adhesive may have better wettability than the resin cement alone and, further, some positive effect (even though low) would be expected from SBU compared with the control group.

Previous studies have stated that resin cement coatings increase the BFS of ceramic materials,^35,41 while silane priming itself does not.²⁵ Based on our results, it cannot be said that silane enhances ceramic BFS, but silane's effectiveness does affect adhesive interface behavior of a ceramic/cement assembly during load and consequently its mechanical properties. This scenario can be extrapolated to a clinical situation in which an indirect all-ceramic restoration is luted with resin cement without treating the internal ceramic surface using an efficient silane primer. In that case, less likelihood of restoration success can be expected, as shown here by m (Figure 2, Table 2). Thus, we may infer that a positive effect of a resin-cement coating on a ceramic/cement assembly's BFS depends on the performance of a proper silanization process. Resin cement coating per se is no guarantee of improved mechanical properties on ceramic/cement systems, as some voids were found across the resin cement layer in all groups (Figures 3 to 6), even though an auto-mixing tip was used, avoiding manual manipulation. Thus, a “perfect” resin cement layer (with no such defects) may be difficult to reproduce in clinical situations, comprising the mechanical reliability of the restoration.

As shown in this study, glass-ceramic/resin cement chemical bonding seems to be crucial not just for bond strength but for a ceramic cement system's mechanical properties, as the integrity of a bilayered system may be considered a more important factor than the strengthening potential of each layer by itself.

CONCLUSIONS

Within the limitations of this in vitro study, the following conclusions can be drawn:

1. The silanization protocol influences the biaxial flexural strength of glass-ceramic/resin cement systems.

2. The application of a silane primer (alone or adhesive coated) to the ceramic surface improves the biaxial flexural strength of glass-ceramic/resin cement assemblies, while the application of a silane-containing universal adhesive does not.