Effect of Resin Cement at Different Thicknesses on the Fatigue Shear Bond Strength to Leucite Ceramic

Abstract Objectives  This in vitro study was performed to evaluate fatigue survival by shear test in the union of leucite-reinforced feldspathic ceramic using different cement thicknesses. Materials and Methods  Leucite-reinforced glass ceramics blocks were sectioned in 2-mm thick slices where resin cylinders were cemented. The samples were distributed in two experimental groups ( n  = 20) according to the cement thickness (60 and 300 μm). The specimens of each group were submitted to the stepwise fatigue test in the mechanical cycling machine under shear stress state, with a frequency of 2 Hz, a step-size of 0.16 bar, starting with a load of 31 N (1.0 bar) and a lifetime of 20,000 cycles at each load step. Results  The samples were analyzed in a stereomicroscope and scanning electron microscopy to determine the failure type. There is no significant difference between the mean values of shear bond strength according to both groups. Log-rank ( p  = 0.925) and Wilcoxon ( p  = 0.520) tests revealed a similar survival probability in both cement layer thicknesses according to the confidence interval (95%). The fracture analysis showed that the mixed failure was the most common failure type in the 300-μm thickness group (80%), while adhesive failure was predominant in the 60-μm thickness group (67%). The different cement thicknesses did not influence the leucite ceramic bonding in fatigue shear testing; however, the thicker cement layer increased the predominance of the ceramic material failure. Conclusion  The resin cement thicknesses bonded to leucite ceramic did not influence the long-term interfacial shear bond strength, although thicker cement layer increased the ceramic material cohesive failure. Regardless the cement layer thickness, the shear bond strength lifetime decreases under fatigue.


Introduction
Dental ceramics are materials routinely used in aesthetic restorative procedures due to their excellent properties, such as high compressive strength, high translucency and fluorescence, chemical stability, low electrical conductivity, and the similar thermal expansion coefficient to the dental substrate. 1he cementation of dental ceramics is a widely studied clinical step, which should be done meticulously to achieve a durable bond strength to the dental substrate. 2,3Thus, resin cements are the first choice materials during the cementation of glass-ceramics restorations, as they provide the advantage of mechanical bonding in addition to the chemical adhesion provided by silanes coupling agent. 4,5][5][6] In addition to the reported parameters, literature shows the influence of resin cement thickness on the fracture resistance of ceramic restorations. 7The authors analyzed the influence of cement thickness and ceramic/cement bonding on the stresses and failures of computer-aided design/computer-aided manufacturing (CAD/CAM) crowns using finite element analysis and monotonic tests.The major findings were that the occlusal "fitting" may promote structural implications in the CAD/CAM crowns; and precementation spaces around 50 to 100μm are recommended. 7ccording to the literature, the benefits of adhesive dentistry are dampened with a cement thickness close to 450 to 500μm due to polymerization shrinkage stresses. 6,7However, a previous report concluded that the cement layer thickness did not interfere in the mechanical performance of ceramic restorations. 8,9However, there is a lack of information in literature regarding the influence of cement thickness on the bond strength durability.
Besides to the previously described, the presence of moisture and chewing loads in the oral cavity are factors capable of impairing the restoration longevity with the degradation of cement layer, slow crack growth, and fatigue. 10The masticatory loads promote mechanical degradation, whereas moisture corrodes the chemical bonds at the ceramic crack tip.In the same manner, moisture from saliva and dentinal tubule can also greatly influence the degradation of resin cements, and finally leads to failure at the adhesive interface. 10However, the survival probabilities regarding the fatigue of ceramic bonded to cements have not been assessed so far.
Therefore, there is lack of scientific evidence in literature showing the influence of mechanical fatigue on the resin cement and glass ceramic bond strength survival.The objective of this study was to assess the bond strength survival of leucite-reinforced feldspathic ceramic (IPS Empress CAD Multi, Ivoclar Vivadent, Schaan, Liechtenstein) during shear stress using different cement thicknesses.The null hypothe-sis was that the different cement thicknesses will not influence the fatigue shear bond strength of leucite-reinforced feldspathic ceramic.

Materials and Methods
First, specimens were made using leucite-reinforced glass ceramics to evaluate the influence of cement thickness on the shear bond to leucite-reinforced feldspathic ceramics, where a resin cylinder was cemented on one of the sides for it to be subjected to the mechanical shear test.The materials used in this study are shown in ►Table 1.

Ceramic Processing
Ceramic blocks (IPS Empress CAD Multi) were obtained and then sectioned in a standardized manner using a precision cutter (Isomet 1000; Buehler, Lake Bluff, Illinois, United States) under constant water cooling to obtain 2-mm thick slices (5 Â 5 mm).A total of 40 slices were obtained, which were polished in a polishing machine with 600 grit sandpaper under constant water cooling.Then, they were randomly distributed into two experimental groups according to the cement thickness used in the cementation of the ceramic (n ¼ 20).

Resin Cylinders Manufacturing
Composite resin cylinders (Opallis, FGM, Joinville, Brazil) (3.2 mm Â 4 mm) were made using a standardized silicone matrix and the incremental technique under a glass slide to obtain a completely smooth surface, and light curing was performed for 30 seconds on each layer with light source unity (Blue Phase, Ivoclar Vivadent) at a light intensity of 1,200 mW/cm 2 .

Standardization of Cement Thickness
To standardize the cement thickness, a pilot study was performed in which the specimens were cemented witch different weights at static load.After the cementation, the specimens were embedded in acrylic resin cylinders (2.5 cm Â 2.5 cm) and cut into halves using a precision cutter (Isomet 1000; Buehler).Then, each sectioned sample was submitted to the microscopy analysis (Discovery V20, Carl Zeiss, Jena, Germany) aiming to define the average thickness generated by each load.In the end, the control group (60 μm thickness) was achieved with a load of 51.2 g and the second group (300 μm thickness) received a higher weight of 811 g (►Fig. 1).Therefore, in the present study one group presented a thin cement layer (60 μm) and the other group presented 5Â this cement layer thickness (300 μm).

Cementation of Specimen
Next, the ceramic surface was isolated with a perforated matrix correspondent to the adhesive area of the cylinders samples to avoid excess of cement (►Fig.2).The isolated surface was etched with hydrofluoric acid 5% (►Fig.2A) (Condac Porcelain, FGM) for 60 seconds, washed with water for 60 seconds, and then a jet of air was used for 15 seconds to dry the surface.The silane (Monobond N, Ivoclar Vivadent) was applied with the aid of a microbrush for 60 seconds (►Fig.2B).After surface treatments, the resin cement (Multilink N, Ivoclar Vivadent) pastes were mixed as recommended by the manufacturer and applied on the center of the treated ceramic surface and on its counterpart on the resin cylinders.The ceramic and the cylinders were bonded together and a static load of 51.2 g weight was applied to the top surface of the cylinder to the 300-μm group; and a load of 810 g was used to the 60-μm group (►Fig.2C).Then, the cement excesses were removed, and light activation (1200 mW/cm 2 , Bluephase N, Ivoclar Vivadent) was performed in five exposures of 20 seconds on each side of the sample.The etched ceramic surface and the polished ceramic can be observed with scanning electron microscope (SEM) in ►Fig.3.

Monotonic Test
The immediate bond strength test was performed to determine the fatigue survival test profiles.To do so, a microshear assay (DL-1000, EMIC, Instron, São Jose dos Campos, Brazil) was performed with a load cell of 20 kgf (0.5 mm/min) in the samples (n ¼ 6) of both groups (►Fig.4A).

Stepwise Fatigue Test
2][13][14] The fatigue test parameters were defined according to the monotonic test results: the frequency was 2 Hz, with step size of 0.16 bar, starting with a load of 31 N (1.0 bar) during 20,000 cycles at each load step until the survival or failure of the sample (suspension).The specimens of each group (n ¼ 20) were submitted to a fatigue test in a mechanical cycler (Biocycle, BIOPDI, São Carlos, São Paulo, Brazil).The uniaxial load was applied with a 12-mm diameter stainless steel piston in the lateral surface of the ceramic, which was supported and arrested between two flat steel bases perpendicular to the load incidence.The test was performed with the samples submerged in water (►Fig.4B).

Failure Analysis
After the fracture, the samples were analyzed in a stereomicroscope (Discovery V20, Carl Zeiss) under magnification of 7.5Â to determine the failure type, and representative specimens (n ¼ 3) from each group have been analyzed with SEM (FEG-SEM, Inspect S50; FEI Company, Brno, Czech Republic) in high vacuum, using a voltage of 25 Kv.Previously, the samples were metallized using the EMITECH SC7620 Sputter Coater metallizer, resulting in a 12-nm gold alloy layer over each sample.Two distinct types of failure were recorded: (1) adhesive failure was recorded when the bond failure was observed in the ceramic and (2) mixed failure was recorded when involving the cement and ceramic surface.The images were observed in secondary electrons and backscattered electrons at low and high magnification (►Fig.5).

Data Analysis
The fatigue data were subjected to a Kaplan-Meier analysis with post hoc Mantel-Cox (log-rank) and Wilcoxon test (α ¼ 0.05) (IBM SPSS Software; IBM, Armonk, New York, United States), as well as a reliability analysis by the Weibull test (Weibull þþ, Reliasoft, Tucson, Arizona, United States).
The mean between monotonic and fatigue test was calculated.The failures were classified and the percentages of each type of failure computed.

Results
The Kaplan-Meier fatigue survival chart and the Weibull probability plots versus number of cycles during the fatigue test are presented in ►Figs.6 and 7, respectively.There is no  significant difference between the mean values of shear bond strength according to both groups comparison.Log-rank (p ¼ 0.925) and Wilcoxon (p ¼ 0.520) tests revealed a similar survival probability in both cement thickness groups (300 and 60 μm) according to the confidence interval (►Table 2).
In addition, the sample failure during the fatigue test increased progressively with the increase in the number of cycles and load (N).The Kaplan-Meier estimates are presented in ►Table 3.No differences in mean values of bond strength (MPa) were observed between the two thicknesses for monotonic and fatigue tested interfaces (►Table 4).
The fracture analysis showed that predominantly ceramic mixed failure was the most common failure type in the 300μm thickness group (80%), followed by adhesive failure (20%).Moreover, the adhesive failure was predominant in the 60μm thickness group (67%), followed by mixed failure (33%).The results (%) are shown in ►Fig.8.The images of failed specimens of each group and the topography micrographs polished and etched of ceramic obtained with the SEM are shown in ►Figs.4 and 5, respectively.

Discussion
The goal was to assess the bond strength survival of leucitereinforced feldspathic ceramic during shear stress using different cement thicknesses.The results showed that the cement thickness did not influence the bond strength survival, thus confirming the null hypothesis.
The leucite-based ceramic (SiO2, Al2O3, K2O, Na2O, other oxides, and pigments) are rich in silica, being considered an acid-sensitive ceramic. 4It is in the glass-matrix structure that the hydrofluoric acid will modify the surface topography and form micro-retentions which improve the bond strength with cement material.Thus, this type of dental ceramic presents a strong bond strength with resinous cements, which endorses our findings.However, the literature reports that the polymerization shrinkage during light curing generates residual stress at the adhesive interface and reduces the adhesion durability of bonded ceramics with a thick cement layers. 15,16Overall, the matter regarding cement thickness has been much explored in fiber post-studies; and a consensus about the association of thinner cement layer and higher bond strengths already exists. 17,18Thus, according to the previous studies, the presence of voids and bubbles are responsible for the decrease of bond strength in thicker cement layers.
However, the well-controlled scenario during cementation and the flat surfaces in the present study probably impaired the inclusion of such defects and no differences were found between 300 and 60 μm cementation layers.
In addition to the defects population on the cement, Fleming et al 19 found that the volumetric shrinkage of the resin cement generates compressive stress at the interface, producing residual stress that allow the crack growth phenomena. 19On the other hand, marginal discrepancy and lower bond strength caused by thicker resinous materials could be more evident in complex geometry, such as crowns,     due to high polymerization shrinkage of the cement. 20,213][24] It was reported that the acceptable cement thickness for dental restorations should be up to 120 μm. 25 Therefore, in the present study one group presented a clinically acceptable thickness (60 μm) and the other present 5Â the cement layer height (300 μm).Despite this difference, the bond strength was similar between both.
One important aspect in bond strength studies is the failure mode evaluation.Adhesive failures occurs when there is no critical damage caused in the ceramic surface, and this factor is particularly important because these results are closer to the real bond strength between ceramic and cement.On the other hand, the thicker cement layer (300 μm) presented predominantly failures involving the ceramic substrate, with chipping of the ceramic surface.This result was caused probably due to the increase in the momentum during the test and the residual stress by the bigger cement volume.This may clinically signify that a cushioning effect between the crown and dentine due to the presence of a low elastic modulus layer (cement) under a much stiffer material (ceramic crown) is potentially lost in the 300-μm cement layer, thus leading to tensile fracture of the ceramic at the surface. 24,25he reliability of both groups was the same and express that the bond strength will diminish over time due to fatigue, and 63.2% of the specimens will fail by around 90,000 cycles.In this study, a mechanical fatigue machine was applied for the aging process, but the specimens were immerged in water which can also contribute to cement plasticization and interfacial degradation.
For the lifetime prediction of a fatigue test to be considered clinically relevant, the evaluated restorations should be evaluated for at least 10 6 cycles.This period of aging would correspond to approximately 1 year of clinical use.Considering three periods of 15 minutes of chewing per day, the individual average of chewing is 2,700 times a day with a frequency of 1 Hz. 26These parameters may vary depending on the applied load and frequency.
The results obtained in our study regarding the survival analyses (Kaplan-Meier) show that more than half of the samples in both groups (300 and 60 μm) failed with approximately 20.40 MPa, and consequently the survival probability drops by a half with every cycle.The reliability plot also showed the same trend, as reliability decreased over time under fatigue.This shows the necessity of further analysis using fatigue tests to predict the long-term behavior of bonded restorations.
Dal Piva et al 27 observed shear stress and tensile stress simultaneously; however, they found high shear stress at the interface of the tested materials than tensile stress.In this study, the shear bond test was performed for monotonic and fatigue tests, in which the monotonic values were used to obtain the fatigue profile and load steps for the survival curve of each thickness group.This methodology was widely applied in studies that evaluated failures in restorative  materials, 12,14,28,29 however, never with the present setup to test shear bond durability.In this study, the cement thickness was standardized by applying a specific amount of weight using an adapted apparatus device for each specimen (►Fig.2A).The study of Ustun and Ayaz 30 used digital pressure when the resin cement was applied to the ceramic surfaces and seated to the dentin, which technically is not a standardizable procedure since there is no measured and control, and different pressures could be applied depending of the operator.A previous study applied different weights using a standardized pressure device, 31 and similar to the present study the authors were able to control the cement layer thickness; however, they also reported the need to check each sample under microscopy prior to the bond strength test due to the technique sensitivity. 31ccording to the results of this study, the different cement thicknesses in the adhesive interface with leucite-reinforced feldspathic did not influence the fatigue behavior during cyclic shear stress state.However, we still advocate the control of cement thickness for better results in ceramic crowns restorations. 32n important limitation inherent from the present study was the load application method.When testing adhesion using shear bond test, a combination of shear and tensile forces occur at the interface, resulting in complex stresses.In addition, the jig design applied in the in vitro setup show variations complicating direct comparison between studies. 33Therefore, the comparisons between different shear bond strength methods should be carefully performed since the stress distributions have nonuniform tensile or shear stress states at the interface due to different specimens' geometries and loading configuration (i.e., wire loop, knife edge, blade, hook, point of loading, alignment, stressing rate) and modulus of elasticity of the evaluated materials. 34Since the present study applied the fatigue loading using a different testing device, the results presented here could have been affected by it, in a similar way for both evaluated groups.
Comparing the loading method for shear bond strength test, the wire-loop, notched-rod, and knife-edge chisel are usually reported as the most preferred blade designs. 35Since the knife-edge chisel transmits the force to the tested sample from a single point, a heterogeneous force transmission is observed.Despite that, the failure mode between the different application methods seems to be similar. 35According to the literature, problems related to the validity of shear bond strength values started to arise as cohesive failures in the substrate were frequently observed with adhesives that yield improved bond strengths. 364][35][36][37][38][39] Therefore, further fatigue bond-strength tests should be developed, assessing not only the reliability of bonded restorations but also the loading method stresses.
Considering the limitations of an in vitro study, such as the use of an accelerated life test and nonanatomic specimens, further investigations could be considered to analyze the influence of the cement thickness in conditions simulating anatomical specimens, as well as the behavior under different surface treatments, different ceramic materials, and varying load profiles in the adhesively cemented restorations.

Conclusion
The resin cement thicknesses bonded to leucite ceramic did not influence the long-term interfacial shear bond strength, although thicker cement layer increased the chances of ceramic cohesive failure.Regardless the cement layer thickness, the shear bond strength lifetime decreases under fatigue.

Fig. 2
Fig. 2 Cementation process with a perforated matrix.(A) Isolated surface etching with hydrofluoric acid 5%.(B) Silane applying with the aid of a microbrush.(C) Cementation of the resin cylinders using static loading.

Fig. 6
Fig.6Survival plot using the Kaplan-Meier method, average and median strength of samples during the fatigue test.

Fig. 7
Fig.7Reliability plot showing that samples will fail as the time under fatigue increases in both groups.The Weibull modulus (β) and characteristic life (η) for 300 and 60 µm groups were respectively: β ¼ 3.2 and 3.3 and η ¼ 94,000 and 90,000 cycles.
Effect of Resin Cement at Different Thicknesses on the Fatigue Shear Bond Strength to Leucite Ceramic Arcila et al. 1321

Fig. 8
Fig. 8 Type of failure data of each group (%) regarding the scanning electron microscope (SEM) analysis.

Table 2
Survival probability in both cementation thickness groups (300 and 60 μm) according to the confidence interval of log-rank (p 0.925) and Wilcoxon (p 0.520) method

Table 3
Results of shear bond fatigue test in MPa, number under risk, surviving specimens, survival probability, standard error, and confidence intervals (CIs) Note: Kaplan-Meier analysis with post hoc Mantel-Cox (log-rank) and Wilcoxon test (α ¼ 0.05).

Table 4
Mean values of bond strength (MPa) obtained in monotonic and fatigue tests