Shear Bond Strength of Lithium Disilicate to Resin Cement After Treatment with Hydrofluoric Acid and a Self-etching Ceramic Primer

aUniversidade Federal Fluminense (UFF), Escola de Engenharia Industrial Metalúrgica de Volta Redonda (EEIMVR) Av. Trabalhadores, 420, V. Santa Cecília, CEP 27255-125, Volta Redonda, RJ, Brasil. bCentro Universitário de Volta Redonda (UniFOA), Avenida Paulo Erlei Alves Abrantes, 1325, Três Poços, Volta Redonda, RJ, Brasil. cInstituto Militar de Engenharia (IME), Departamento de Ciência de Materiais, Pr Gen Tibúrcio, 80, Praia Vermelha, CEP 22290-270, Rio de Janeiro, RJ, Brasil. dUniversidade do Estado do Rio de Janeiro (UERJ), Faculdade de Tecnologia de Resende (FAT), Rodovia Presidente Dutra, km 298, CEP 27537-000, Resende, RJ, Brasil.


Introduction
The development of new glass-ceramics with good mechanical properties combined with aesthetics results, allowed the manufacture of metal-free ceramic dental prostheses for the restoration of different tooth elements 1-3 . CAD-CAM (Computer-Aided Design / Computer-Aided Manufacturing) has several advantages over conventional manufacturing of prostheses, such as the ability to prepare the restoration in less time, uniform morphology, automation of the manufacturing process and better mechanical properties 4 .
CAD-CAM blocks are available for aesthetic restorations, including yttrium tetragonal zirconia polycrystals, Y-TZP 5,6 , feldspathic glass ceramics 7,8 , lithium silicate glass ceramics 9,10 , composite resins or hybrid ceramics 11,12 . The ceramic performance of adhesive cementation is influenced by variations in the chemical composition of these materials, which modify their mechanical properties 13 .
Lithium disilicate glass-ceramic (Li 2 Si 2 O 5 ) has excellent optical properties and mechanical resistance, is used in many dental applications, such as veneers, inlays, onlays, single and full crowns 3,14 . This biomaterial has adequate mechanical properties to be used in dental restorations with thicknesses up to 0.5 mm and the results are equivalent to those obtained with zirconia covered with porcelain 6 . Li 2 Si 2 O 5 is available in blocks for CAD/CAM systems. After manufacture, a thermal treatment is necessary to promote the phase transformation to of lithium disilicate and a reduction of the volume fraction of amorphous material. After this procedure the flexural strength is about 400 MPa 10 .
The mechanical properties of materials used in dentistry (ceramic restorations, resin cements, silane coupling agents, tooth structures) and the surface optimization of bonding *e-mail: claudineisvr@gmail.com substrates influence the performance of prosthetic restorations.
To have a good ceramic adhesion, it is essential to prepare the tooth structure and surface restoration. Insufficient bond strength between restoration and resin cement promotes an inhomogeneous stress distribution at the bonding interface, inducing the failure of the cement-restoration bond under the action of masticatory forces [15][16][17] .
The surface treatment before cementation is very important for the success and longevity of ceramic restorations. In this treatment, a micro-roughness is created on the surface of the ceramic to promote bonding to the resin cement 18 . The protocol treatment varies according to the material used. Feldspathic glass-ceramic and leucite-reinforced glassceramic, for example, are treated for 60 s with hydrofluoric acid (5 -10% HF), while lithium disilicate requires only 20 seconds 17,19,20 . However, ceramics based on alumina (Al 2 O 3 ) and zirconia (ZrO 2 -Y 2 O 3 ) have high chemical stability and are commonly treated with airbone particle abrasion before adhesive cementation.
The protocol for treatment of glass ceramic is HF acid etching to create surface roughness. In the treatment, the glass matrix is selectively removed by controlling the time of exposure to a fatty acid with lower chemical than the Li 2 Si 2 O 5 crystals. The roughness created with HF etching is suitable for micromechanical retention on the ceramic surface 21 . This rough surface increases the surface energy and facilitates the subsequent treatment with silane. Silane has a bi-functional molecule, in which one end reacts with the glass phase of the ceramic surface and the other end copolymerizes with methacrylate groups in the organic matrix of resin cements through siloxane links. The mechanisms involved characterize adhesive cementation 22,23 . The use of silane in dentistry has been successful, improving the bond strength of the ceramic with the resin cement 24,25 . However, severe surface changes can decrease the flexural strength of materials due to surface defects that induce the formation of cracks. It is thus important to control the concentration and conditioning time to obtain surface changes without weakening the ceramic 22,24,26,27 . If the treatment is performed in the intraoral cavity, care must be taken in relation to the high toxicity of HF-solution, which can generate necrosis of the soft tissues and bones 28 . In addition, once one cannot use it in the mouth, an alternative treatment is desirable. These two deleterious mechanisms depend on the time of exposure and the concentration of the HF solution.
A self-etching ceramic primer 29 has been used as an alternative to the traditional HF acid etching procedure [30][31][32][33] . A self-etching ceramic primer is basically composed of ammonium polyfluoride 34 , trimethoxypropyl methacrylate, solvents (alcohol and water) and a green pigment to provide visibility) in an all-in-one system (one-step etching technique). This procedure requires less time, does not weaken the glass ceramic because it is less aggressive and avoids the toxic risks of hydrofluoric acid 17,24,29,[35][36][37] .
The aim of this study was to evaluate the effect of the use of a self-etching ceramic primer (SECP) in the treatment of the surface of lithium disilicate samples, to determine the shear bond strength to resin cement and to compare the results with the conventional HF acid etching technique. The microstructural and crystallographic aspects of lithium disilicate after heat treatment and its correlations with resistance to adhesion with resin cement are discussed.

Materials
Lithium metasilicate (Li 2 SiO 3 ) blocks used in the manufacture of dental prostheses were transformed by a heat treatment (840 °C -7 min) into lithium disilicate (Li 2 Si 2 O 5 ) and, the surface was prepared for a bonding strength test with resin cement. The specifications of all materials are shown in Table 1.

Processing
Commercial blocks of lithium metasilicate IPS E-max CAD CAM (Ivoclar Vivadent Schaan, Liechtenstein) with dimensions of 14 x 18 x 12 mm were cut into approximately 10 x 10 x 3 mm plates using an Isomet-Buehler 1000 cutter (Buehler -Germany). The plates were ultrasonically cleaned with distilled water for 15 min and subjected to a heat treatment at 840 °C for 7 min under vacuum for phase transformation. The heat treatments were carried out in an Ivoclar P 5000 furnace (Ivoclar Vivadent Schaan, Liechtenstein), following the manufacturer's recommendations. After the phase transformation, the samples were polished with Al 2 O 3 suspension, to standardize surfaces and facilitate comparative analysis, using Aropol polisher (Arotec, São Paulo -Brazil).

Characterization
Before and after the heat treatment, the samples were characterized by X-ray diffraction using an XRD-6100 diffractometer (SHIMADZU Corp. Japan), aiming to determine phase transformations that will occur in the investigated glass-ceramic during the dental prosthesis manufacturing protocol. Fragmented samples (particle size < 32 μm) were analyzed using Cu-Kα radiation, in the 2θ range from 20 to 90°, using a step width of 0.05° and a counting rate of 5 s/step. The XRD patterns were compared with those of the Crystalographica Search-Match software (Oxford Cryosystems). Phase quantification was performed using Rietveld refinement with the FullProf Suite 3.0 software.
Polished glass-ceramic bulk samples, before and after the heat treatment were imaged in a scanning electron microscopes Zeiss EVO LS15. In order to reveal the microstructural features, the polished surfaces were etched with a 5% HF solution for 20 s. After acid etching, the samples were coated with a thin layer of gold using an Emitech K550X Sputter Coater.

Sample preparation
To analyze the influence of the surface treatment on the shear strength to resin cement, the samples were divided in three groups ( After the lithium disilicate surface polishing and HF acid etching or treated with self-etching ceramic primer, the surface roughness of the samples was measured using a Zygo New View 7100 Optical Profiler. The parameters of the roughness Ra, PV and Rz were measured. The test was performed according to the recommendations of the ISO 4288:2008 38 and ISO 4287:2002 39 standards. These surfaces were coated with a thin layer of gold and analyzed by SEM using a microscope model EVO-MA10-Zeiss. Figures 1 and 2 show the different sample preparation steps to determine the shear bonding strength of lithium disilicate to resin cement. Before acid etching (10% HF) and SECP, the lithium disilicate samples were embedded in acrylic resin and covered with a 2 mm thick silicone (Express 3M) plate, with four perforations with a diameter of 2 mm, which were filled with Allcem Core (FGM) dual resin cement. After removing excess resin cement, the samples were light-cured for 40 s, using an Ultradent VALO Cordless LED light curing device, simulating the clinical practice. The silicone mould was cut with a scalpel and removed from the sample. A second light cure for 40s was performed on the resin cylinders with the same light curing device. The samples ( Figure 1) were stored in a container with 70% relative humidity. The shear strength test was made 24 h after sample preparation and bonding.
The cross sections of the glass-ceramic/resin cement interfaces were analyzed by SEM using the microscope model EVO-MA10-Zeiss. In this stage, the analyzed surfaces Table 1. Technical specifications and main characteristics of the materials used (data from the manufacturers).

Material
Code With the aid of a disposable microapplicator, apply Ambar Universal to the internal surface of the part (previously treated) under friction, wait 15s and then apply a light jet of air for 10s. Non-light-curing were also covered with a thin layer of gold to make them conductive.

Shear bond strength testing
The shear bond test was performed using the device shown in Figure 3, following procedures of previous works 40 , using an EMIC DL10000 universal testing with a 20 N load cell. The loading speed was 1.0 mm/min.
The shear strength of the interface between lithium disilicate and resin cement was calculated using the equation where σ shear is the shear stress in MPa, F is the maximum failure load in N and A is the adhesion area in mm 2 .

Statistical analysis
One-way ANOVA analysis was used to evaluate the mechanical properties, followed by a Tukey's Honestly Significant Difference (HSD) post hoc test (α = 0.05) to determine the difference between the means of roughness and shear strength. The statistical analysis was performed with the software ASSISTAT version 7.7 beta.
The complementary statistical evaluation of shear strength was carried out using Weibull statistics 41 where P is the failure probability, m is the Weibull modulus, σ 0 is the characteristic stress in MPa and σ is the average bonding strength in MPa.    responsible for by the halo observed in the XRD pattern, Figure 4a. After the heat treatment, Figure 4b, the percentage of Li 2 Si 2 O 5 increased to approximately 72.6% with 5.6% of Li 2 SiO 3 and 21.4% of the amorphous phase. Figure 5 shows typical SEM micrographs of the sample surface before and after heat treatment. One can see in Figure 5a a large amount of amorphous phase surrounded by equiaxial crystals of lithium metasilicate (Li 2 SiO 3 ) before the heat treatment. These observations are consistent with the XRD results. After the heat treatment, as shown in Figure 5b, there are elongated and interlaced grains of Li 2 Si 2 O 5 surrounded by the residual amorphous phase partially extracted by the HF-solution chemical etching. Table 3 and Figure 6 show the results of the roughness tests of the samples. Table 3

Shear strength and statistical analysis
The results of shear strength and statistical analysis are shown in Table 4. Figure 8 shows the Weibull distributions of the shear tests. Groups NT, HF and SECP had shear strengths of 5.31 ± 0.94 MPa, 24.25 ± 1.21 MPa e 24.80 ± 1.82 MPa, respectively. Statistical analysis showed that there was no significant difference between groups HF and SECP. These groups have significant differences in shear strength from control group (NT).
The Weibull moduli of HF group (m HF = 19) and SECP (m SECP = 13) were higher than the control group NT (m NT = 6). The lower value of the Weibull modulus of the group without surface treatment (NT) indicates greater dispersion of shear strength, which is associated with a lack of chemical and / or mechanical adhesion, between the lithium disilicate and the resin cement, allowing the parts involved in the interface, and consequently in the shear strength, are strongly associated with the surface characteristics and the standardization of the surface preparation protocol. These results confirmed change in topography and in the surface roughness of the ceramic material due the surface treatment with 10% HF. The HF etching increased the micromechanical interlock and the bond strength between the resin cement and the glass ceramic. The self-etching ceramic primer promotes changes in the surface of the lithium disilicate sample and adhesive strength similar to Group HF even with a less pronounced conditioning pattern.
Figures 9a-c shows the resin cement-lithium disilicate interfaces for the three groups of samples. It is observed in Figure 9a (NT, control group) the presence of empty spaces at the interface, showing a deficient bonding between the resin cement and the lithium disilicate ceramic. There is no resin cement infiltration and heterogeneous mechanical    interlock. Figures 9b and 9c show a homogeneous layer on the cement-glass-ceramic interface after hydrofluoric etching (HF) and treatment with the self-etching ceramic primer (SECP). The irregularities caused by HF etching or SECP and mechanical interlock were completely filled out.

Discussion
The phase transformations of the lithium silicate, resulting from the heat treatment proposed by the manufacturer (840 °C-7min), are well known. and studied 42 . Based on the chemical composition of the glass informed by the manufacturer (SiO 2 :60-80% Li 2 O:11-19%, K 2 O:13%, P 2 O 5 :11%, ZrO 2 :0-8%, ZnO:0-8%, Al 2 O 3 : 0-5%, MgO: 0-5%, and pigments), and from the XRD analyzes performed on the as-received lithium metasilicate, Figures 4a and 4b, it can be inferred that the formation of Li 2 Si 2 O 5 crystals that occur at 840 °C -7min, come mainly from two simultaneous chemical reactions, and are based on the presence of Li 2 SiO 3 crystals (metastable above 700 °C), and the presence of 54.5% of residual amorphous phase, rich in SiO 2 and Li present in Li 3 PO 4 nuclei dissolved in this residual glass of the crystallized samples.
As identified in Figure 4c, the crystallization occurred at 840 °C-7min. In addition to elongated crystals of Li 2 Si 2 O 5 , Figure 5b, the heat-treated material presents an amount of residual amorphous phase in the order of 21%. The microstructural aspects indicate that the grains of Li 2 Si 2 O 5 are elongated (high aspect-ratio) and with an average size of 1 to 2 μm, which increases the atomic roughness of the surface of these crystals. This information is important, since this 21% of residual glass, basically contains silica (SiO 2 ) in addition to ions of such as K, P, Al, Zn, Mg. The silica matrix present in the residual glass is responsible for the activation of chemical reactions that allow chemical bonding and anchoring between the glass ceramic matrix and the resin cement, and the increase in the atomic roughness of the elongated crystals helps in the anchoring between the two layers.
The lack of these chemical and topographic interactions between the surfaces are the main reasons for lowering the bond strength presented by the control group (NT), in comparison to the HF and SECP groups.
The cementation defects of dental prostheses occur at the ceramic-resin interface. Adhesion plays an important role in long-term success, since it improves marginal adaptation of the restoration, decreases microleakage and increases fracture resistance when the prosthesis is subjected to masticatory loads 14,15 . The acid etching surface treatments of ceramics dental prostheses before cementation creates micro-porosities due the removal of the glass phase. Microporosities increase the surface area, increase the surface energy and favour chemical bonding. The bonding agent during the silanization process improves the adhesion. Once adhesive cementation is completed, the formation of a tooth-restoration monoblock is sought to increase the longevity and durability of indirect adhesive restorations [23][24][25]43,44 .
The 3D roughness analysis ( Figure 6) showed that the surface treatment increased the roughness of the samples relative to the control group. This result is consistent with the SEM micrographs ( Figure 7) and previous results reported in the literature 17,24,35,45 . HF etching acts dissolves the amorphous component, exposing the lithium disilicate crystals and yielding a rough surface suitable for cementation. In this work, HF 10% treatment was performed for 20 s, followed by water jet removal, but one must be careful, because small increases in the treatment time and/or HF concentration can promote significant changes in the microstructure that are detrimental to its clinical performance.
The surface treatment with the one-step self-etching ceramic primer yields a smoother conditioning pattern ( Figure 6). Tetrabutylammonium dihydrogen trifluoride in SECP is Table 3 shows that the SECP treated samples have a surface profile with shallower valleys and fewer stress sites than HF treated ones, as also observed by Tribst et al. 45 . This morphology improves the mechanical behavior under cyclic loading of lithium disilicate. In a practical way, SECP is a simple process for glass-ceramics surface treatment that decreases damage to the microstructure of adhesive restorations, improve mechanical interlocking and provides enough adhesive resistance to t minimize the risk of handling accidents by the dentist and his assistants The conventional surface treatment protocol is HF etching and silanization. Silanes are mediators and provide stable resistance between resin cement and glass ceramic. The silane used in this work is 3-methacryloxypropryltrimethoxysilane, a bifunctional molecule with an organofunctional group containing methyl methacrylate that copolymerizes with resin cements and hydrolyzable alkoxyl groups (silanol) that react with Si-OH on the ceramic surface 24 . The alkoxyl group activates the condensation reaction with the hydroxyls present in the residual amorphous phase and in the glass ceramic, releasing water and products. The 3-methacryloxypropryltrimethoxysilane forms siloxane bonds resulting in a cross-linking tri-dimensional layer. The formation of this layer results in the adhesion of resin cements 34,44 .
The adhesive bond strongly depends on the irregularities created in the ceramic surface, which increases the free surface energy and improves the dispersion of silane and resin cement. The 3-methacryloxypropryltrimethoxysilane molecules present in the silane bind to the hydroxyl groups in the ceramic and decreases surface energy 34 . According to Prado et al. 24 , cleaning with water and drying with air after applying SECP leaves a thin layer of silane on the ceramic surface, promoting the removal of ammonium polifluoride and the reaction by products forming siloxane bonds. The results of the shear test show an increase in the bond strength in samples treated with HF acid (Group 2) and self-etching primer (SECP) relative to those of the control group (NT), in which there is no considerable mechanical interlocking or silanization. In this study, the bond strength of SECP treated samples is statistically similar to that of HF treated ones.
Although the results of shear strength are statistically similar for groups HF and SECP, with average values of 24.25 MPa and 24.8 MPa respectively, the Weibull moduli are different the three groups ( Figure 8). The group without conditioning (NT) had a Weibull modulus m NT = 6, which shows the scattering of results due to the lack of a uniform adhesive layer. Groups HF and SECP have Weibull moduli of m HF = 19 and m SECP = 13, respectively. The difference in terms of reliability and spreading of the results reveals that the HF group has less bonding variation among samples. This result is due to the efficient creation of surface roughness and increased surface area. HF etching improves glass phase dissolution, increases the wettability of silane that fills in the roughness and promotes a homogeneous surface layer, but also improves the wettability of the adhesive system.
The combination of acid and silane in a single bottle brings advantages such as saving time, decreasing the risks of acid in handling and avoiding the weakening effect on ceramics due to the action of ammonium polyfluoride, which promotes a less severe dissolution of the glass phase of the ceramic, while silanizing the conditioned surface in situ 24,34,[45][46][47] .
According to Moreno et al. 34 , positive results were obtained with the use of SECP. The self-etching ceramic primer was able to transform the hydrophilic surface of the ceramic into a hydrophobic substrate by decreasing the free surface energy and the contact angle between the liquid and substrate. Contaminant removal increases cross-linking in the silane layer and the amount of reactive binding sites available. SECP produced a highly hydrophobic silane layer that favours the hydrolytic stability of the cement-resin interface, forms siloxane bonds with the ceramic surface of the lithium disilicate and improves the penetration of the resin cement by promoting mechanical interlocking. The analysis of the bonding interface in in vitro studies carried out by Murillo-Gomez et. al. 36 showed similar performances of SECP and acid conditioning. The authors attribute this similar efficiency to the silane contained in SECP, which forms a water-resistant layer, providing a chemical bond between the glass phase and the resin cement. The single-stage primer cleaning step appears to remove the contaminants left by the acid on the surface better than the air jet after applying silane separately in the standard protocol because, due to chemical affinity, water eliminates such by products.
The pattern shown in the SEM micrographs of the cross sections, of the bonding interface between resin cement and lithium disilicate glass-ceramic, Figure 9a-c, are consistent with the similarity of statistical results for both techniques. The bonding interface formed with conventional and simplified surface treatment proved to be homogeneous, showing that there was complete filling of the irregularities created by conditioning for both techniques, promoting mechanical interlocking between resin cement and ceramic. This condition has a positive influence on the quality of the bond and is in accordance with the values obtained in the bond strength tests. The ceramic-resin interface after application of silane and adhesive was deficient because the resin cement did not penetrate the ceramic surface, as shown in the image (Figure 9a) the unfilled spaces. The unfilled spaces generate areas of stress concentration leading to the failure of adhesive cementation and decreasing the longevity of adhesive ceramic restorations.

Conclusions
Based on the results obtained in the present study and within the inherent limits of the experimental techniques used, one comes to the following conclusions: The surface treatment with hydrofluoric acid etching (HF) yielded higher roughness parameters than the control group without surface treatment (NT) and the self-etching ceramic primer group (SECP). Furthermore, the HF and self-etching ceramic primer surface treatments increased the bond strength between the cementing agent and the ceramic surface. The simultaneous action of conditioning and silanizing provided by the selfetching ceramic primer promoted micro-retention and adhesive resistance with minor topographic changes in the microstructure, offering a simplified technique without the risks of hydrofluoric acid.