Introduction The objectives of this study were to determine the effects of various surface conditioning methods on 3 types of ceramic materials (feldsphatic, leucite-based, and lithia disilicate-based) in orthodontic bonding. Methods A total of 210 ceramic disk samples were fabricated and divided into 3 groups. In each group, 5 subgroups were prepared by sandblasting; sandblasting and hydrofluoric (HF) acid; sandblasting and silane; sandblasting, HF acid, and silane; and tribochemical silica coating and silane. Mandibular incisor brackets were bonded with light-cured adhesive. The samples were stored in water for 24 hours at 37°C and then thermocycled. Shear bond tests were performed, and the failure types were classified with adhesive remnant index scores. Results In all 3 ceramic groups, the lowest shear bond strength values were found in the sandblasted-only samples. For the feldspathic and lithia disilicate-based ceramic, the highest bond strength values were obtained with silica coating (15.2 and 13.2 MPa, respectively). For the leucite-based ceramic, HF without silane produced the highest bond strength value (14.7 MPa), but comparable values were obtained with silicatization also (13.4 MPa). Conclusions The silica-coating technique could replace the other conditioning techniques in bonding brackets to ceramic. However, debonding must be done carefully because of the risk of porcelain fracture.
With more adults seeking orthodontic treatment, orthodontists must often bond attachments to teeth that have various types of restorations, such as amalgam, gold, composite, and porcelain. Although ceramics are commonly used as esthetic restorative materials, the bond strength of composite resins to ceramic restorations has often been reported to be insufficient. The difficulty for clinicians is that the porcelain surface does not bond readily to other materials. Therefore, ceramic surfaces must be pretreated to achieve sufficient bond strength. Combinations of various mechanical and chemical conditioning methods have been suggested to alter the surface characteristics of porcelain.
The purpose of mechanical alteration of the porcelain is to remove the glaze and roughen the surface to provide sufficient mechanical retention for the orthodontic adhesive. This alteration has been achieved by sandblasting or with a coarse diamond stone.
However, the bond strength of composite resins bonded to ceramic restorations was found to be unsatisfactory in some studies. Furthermore, mechanical roughening with diamond burs and sandblasting causes irreversible damage to the porcelain glaze.
For chemical conditioning, hydrofluoric (HF) acid and acidulated phosphate fluoride can be used. Also, the effect of laser irradiation on the porcelain surface was studied as an alternative conditioning technique. Etching with HF acid produced bond strengths similar to or higher than etched enamel, but, due to its deleterious effect on soft tissues, clinicians must be cautious. The failure rate on ceramic surfaces is still reported to be as high as 9.8% even when HF acid is used before silane application. Silane is a coupling agent that can be used to enhance bond strength to porcelain surfaces. Silane molecules, after being hydrolized to silanol, can form polysiloxane networks or hydroxyl groups to cover the silica surface. Monomeric ends of silane molecules react with the methacrylate groups of the adhesive resins by free radical polymerization. Silanization also increases the wettability of the ceramic surface.
Although sufficient bond strength between metal and porcelain can be achieved, many attempts have been made to develop better techniques for bonding adhesive resin to ceramics. Guggenberger introduced the Rocatec system, based on tribochemical application of a silica layer by sandblasting. After this extraoral system, silica coating was performed with intraoral sandblasting in several studies. In this system, surfaces are blasted with aluminum oxide particles modified with silicic acid. After the tribochemical silica coating, a layer of small silica particles remains on the surface and improves chemo-mechanical bonding via silane application, which provides covalent bonding between a silica-coated substrate and composite resins. Silica coating has been used in restorative dentistry but not thoroughly evaluated for orthodontic purposes.
Ceramic material is believed to have the most esthetically pleasing result for the replacement of a lost tooth or a damaged tooth with an unattractive enamel surface. The demand for better restorations has caused the development of more advanced porcelain systems. There are several types of porcelain for ceramic restorations: silica-based ceramics, glass-infiltrated or densely sintered aluminum oxide ceramics, and zirconium oxide ceramics. Bonding to traditional feldspathic ceramic is a predictable procedure with reliable results when certain guidelines are followed. However, the composition and the physical properties of leucite-reinforced or lithia disilicate-based ceramics are different from feldspathic porcelain and might require alternative bonding techniques.
Our objectives in this study were to investigate the shear bond strengths of metal brackets bonded to 3 porcelain surfaces after various surface-modification techniques, to identify the failure types after debonding, and to determine the effects of various surface conditioning methods on different porcelain surfaces by using scanning electron microscope (SEM).
Material and methods
A total of 210 disk samples, 5 mm thick and 8 mm in diameter, were fabricated from 3 types of ceramic materials: feldsphatic porcelain, leucite-based ceramic, and lithia disilicate-based ceramic. Each ceramic group consisted of 70 samples divided equally into 5 subgroups (each, 14 samples) according to surface-treatment technique.
Feldspathic porcelain (IPS d.sign, Ivoclar-Vivadent, Schaan, Lichtenstein) at a thickness of 2 mm was fired onto the metal disks. Leucite-based ceramic (IPS Empress, Ivoclar-Vivadent) and lithia disilicate-based ceramic (IPS Empress 2, Ivoclar-Vivadent) samples were obtained without alloy bases according to the manufacturer’s recommendations. All samples with glazed surfaces were pretreated with 1 of 5 surface treatment methods before bracket bonding. Table I summarizes the surface conditioning methods and the manufacturers’ names.
||50-μm aluminum trioxide, 1 cm, 3 seconds||
||9.6%, 2 min||Pulpdent porcelain etch gel, Pulpdent Corp, Watertown, Mass|
||30 seconds||ESPE-Sil, 3M ESPE, Seefeld, Germany|
||30-μm silicon-modified A1 2 O 3 , 1 cm, 5 seconds||Cojet-Sand, 3M ESPE, Seefeld, Germany|
Sandblasting with 50-μm aluminum trioxide (GAC, Bohemia, NY) was done vertically from a distance of 10 mm for 3 seconds with an intraoral sandblaster (Microetcher II, Danville Materials, San Ramon, Calif). HF acid (9.6%) (Pulpdent porcelain etch gel, Pulpdent Corp, Watertown, Mass) was applied for 2 minutes, and then washed and dried.
Silica coating was achieved by using the intraoral sandblaster filled with 30-μm silica-modified aluminum trioxide (Cojet-Sand, 3M ESPE, Seefeld, Germany), from a distance of 10 mm for 5 seconds. Silane (ESPE-Sil, 3M ESPE) was applied to the conditioned surfaces and allowed to dry.
The bonding procedures were carried out by the same operator (S.K.). A total of 210 metal mandibular incisor brackets were bonded to the conditioned surfaces with Transbond XT (3M Unitek, Monrovia, Calif). The excess resin was removed, and the adhesive resin was light cured for 40 seconds with a conventional halogen lamp (Ortholux XL 3000, 3M Unitek). All samples were stored in water for 24 hours and thermocycled 500 times between 5°C and 55°C. The shear bond test was performed with a universal testing device (Testometric M500 25 kN, Rochdale, United Kingdom). The sample and the blade were secured to the testing device so that the bonded surface of the porcelain was parallel to the direction of the debonding force. Shear force was applied to the porcelain-bracket interface with a crosshead speed of 1 mm per minute until debonding occurred. The bond strength values were calculated by dividing the maximum load (in newtons) by the base area of the bracket to convert the results to megapascals (N/mm 2 ). The relationship between bond failure and the load applied was shown by the survival analysis. These distribution curves are the cumulative probability of bond failure plotted against the applied shear debonding force for each group and each surface treatment method.
After debonding, the fracture sites were examined to determine the location and classified according to the adhesive remnant index (ARI). For the surface examination, separate porcelain samples were prepared and conditioned as described above. These samples were then sputter-coated with gold and palladium and examined with SEM (Jeol JSM 5200, Tokyo, Japan).
For the statistical analysis, the means of each group were analyzed by 1-way analysis of variance (ANOVA), and the Tukey test was used for multiple comparisons.
Mean bond strength, minimum and maximum values, standard deviations, and standard errors for each group are shown in Table II . In the feldspathic porcelain group, the highest bond strength values were obtained with silicatization (15.2 MPa), HF without silane (11.3 MPa), and sandblasting with silane (10.7 MPa). These values were significantly higher than those of the samples treated by sandblasting only (3.2 MPa). For the leucite-based ceramic, HF without silane (14.7 MPa), sandblasting with silane (12.3 MPa), and silica coating (13.4 MPa) produced similar bond strength values. In this group, significantly lower bond strength values were obtained with HF acid with silane (9.9 MPa) and sandblasting (3.9 MPa). For the lithia disilicate-based ceramic, silicatization (13.2 MPa), sandblasting with silane (11.8 MPa) and HF without silane (8.6 MPa) showed significantly higher bond strength values than did HF acid with silane (5.7 MPa) and sandblasting alone (3.1 MPa). These results were confirmed by survival analysis, which relates probability of bond failure to the load applied. The survival data and the failure rates according to surface conditioning techniques and porcelain types are shown in Figure 1 .
|SB + HF||11.3||1.7||20.4||4.1||1.1|
|SB + HF + S||10.5||2.3||20.5||6.0||1.8|
|SB + S||10.7||1.9||17.2||5.1||1.5|
|SC + S||15.2||3.7||21.5||5.9||1.7|
|SB + HF||14.7||6.2||24.6||5.8||1.8|
|SB + HF + S||9.9||3.6||19.4||5.0||1.5|
|SB + S||12.3||1.1||22.4||8.5||2.5|
|SC + S||13.4||1.8||21.3||6.5||1.9|
|Lithia disilicate- based ceramic|
|SB + HF||8.6||1.3||15||4.8||1.3|
|SB + HF + S||5.7||1.2||14.3||3.6||1|
|SB + S||11.8||2.7||19.7||6.1||1.8|
|SC + S||13.2||0.6||20.4||7.7||2.2|
Table III gives the modes of failure after debonding. In the feldspathic porcelain samples treated with sandblasting and HF acid, the brackets failed mainly at the ceramic/resin interface with all adhesive remaining on the bracket base. With the other conditioning techniques, the brackets failed mainly within the adhesive resin. For the leucite-based and lithia disilicate-based ceramics, adhesive failures between the ceramic and the composite resin were observed mainly in the sandblasting HF acid with and without silane groups. In the sandblasted with silane and the silica-coated groups, the luting cement was mainly debonded from the bracket base and left on the ceramic surface. Cohesive failures in the ceramic were observed more frequently in the samples treated with sandblasting and silane.
|SB + HF||12||2||—||—||—|
|SB + HF + S||6||2||3||—||3|
|SB + S||2||6||2||—||4|
|SC + S||—||3||10||—||1|
|SB + HF||8||3||—||—||3|
|SB + HF + S||10||—||3||—||1|
|SB + S||3||—||—||8||3|
|SC + S||2||—||3||6||3|
|Lithia disilicate-based ceramic|
|SB + HF||12||2||—||—||—|
|SB + HF + S||12||1||—||—||1|
|SB + S||3||—||1||6||4|
|SC + S||2||—||2||8||2|
With SEM, the values obtained from the shear bond testing can be explained further. The SEM photographs showed that sandblasting with 50-μm aluminum oxide modified the morphologic surfaces of feldspathic, leucite-based, and lithia disilicate-based ceramics ( Fig 2 ). The resulting surface topographies were similar between the groups with superficial irregularities and shallow erosion, but the leucite-based ceramics had more microporosities. Silica-coating procedures generated sharply demarcated and acute-angled surface roughness ( Fig 3 ). However, the surface features of leucite-based ceramic ( Fig 3 , B ) are a little different from those of other types of ceramics ( Fig 3 , A and C ), with more pores and shallow erosion. Chemical etching with HF acid showed prominent undercuts and honeycomb-shaped surface irregularities in the leucite-based and lithia disilicate-based ceramics ( Fig 4 , B and C ). The feldspathic porcelain samples, on the other hand, had numerous uniformly distributed pores and grooves between the pores ( Fig 4 , A ).
Our aims in this study were to determine whether there is a difference in bond strength among 3 types of ceramics and to find the most reliable method for bonding metal brackets to different type of ceramic crown materials.
It has been suggested that clinically adequate bond strength for a metal orthodontic bracket to enamel is 6 to 8 MPa. Even though the clinical relevance of in-vitro studies is considered to be limited, the mean shear bond strengths of metal brackets to ceramic surfaces in this study generally exceeded acceptable limits (except for sandblasting) and therefore can be considered sufficient for clinical situations. Among the treatment methods tested, sandblasting with aluminum trioxide only produced the lowest bond strength values, considerably below the acceptable range; therefore, it should not be used as a conditioning technique for porcelain bonding.
Thermocycling is required to test the bond strength of brackets to ceramics, because of artificial aging and the various thermal expansion coefficients of metal, resin, and ceramic materials. Storage in water for 24 hours is normally sufficient to distinguish materials that can and cannot withstand a wet environment. However, thermocycling between 5°C and 55C° is usually used as an accelerated aging test. The temperature changes between the water baths could contribute to water contamination at the resin bond interface and weaken the resin. It was concluded that thermocycling has a significant negative effect on bond strength values; in studies when the samples were thermocycled, bond strength values were lower than those in studies in which thermocycling was not used. Thermocycling of 1000 or 2000 cycles might be recommended by some authors, but, however, 500 cycles is accepted as the ISO standard for adhesion testing.
Shear bond testing after orthodontic bonding and thermocyling has been the standard method of measuring the bond strength of brackets to various surfaces. However, its effect must be questioned because of nonuniform stress distribution, which can cause cohesive failures. Also, the variation of the direction of the debonding force influences bond strength measurements. It was reported that the direction of the debonding force significantly influences shear bond strength values; standardization of this parameter is required for orthodontic shear bond strength testing. In our testing method, all efforts were made to standardize the direction of the debonding forces.
Although the bond strength obtained with HF acid etching was satisfactory, some orthodontists might hesitate to use it because of its potential harmful effects and might prefer a safer alternative conditioning technique. In this study, comparable bond strength values were achieved with HF acid etching and silicatization for the 3 types of ceramics.
There are contradictory reports in the literature on silane application. Some studies reported superior bond strength by using HF acid with silane, but our findings and those of others do not agree. In our study, silane after HF acid application did not increase shear bond strength. This might be due to differences in our testing method. Furthermore, it was shown that the silane completely filled the surface porosity created by HF acid on the lithia disilicate-based ceramic surface. In our study, a significant increase was achieved when silane was applied after sandblasting. But cohesive failures in the ceramic material were observed more frequently in the samples treated with sandblasting and silane.
For the leucite-based and lithia disilicate-based ceramic, more frequently in the silica-coated and the sandblasted and silane groups, the resin predominantly adhered to the ceramic surfaces after debonding (ARI scores of 2 or 3). Apparently, the chemical bond strength was equal to or exceeded the mechanical lock provided by the bracket base, and all adhesive remained on the ceramic surfaces. This type of failure means that the bond strength to the leucite-based or the lithia disilicate-based ceramics was greater than the cohesive strength of the luting resin. The groups treated with HF acid had ARI scores of 0, implying that the bond strengths to the ceramic surfaces were lower than the cohesive strength of the adhesive resin. Sometimes, ARI scores do not correlate with bond strength values ; this was true in this study. The groups with the highest bond strength values had different ARI scores, so one cannot estimate the site of bond failure by a definite shear force. This difference might result from the different bonding types that were obtained with HF acid (mechanical) and silica coating (chemo-mechanical).
Cohesive failure in the ceramic material could indicate that the bond between the adhesive resin and the ceramic was stronger than the ceramic itself. In the silica-coated HF acid-etched or the sandblasted and silane groups, bond strength values reached or exceeded the reported optimum bond strength values but caused cohesive fractures in ceramic restorations. Thurmond et al reported that, when bond strength values between the ceramic and the composite resin exceeded 13 MPa, there would be cohesive fractures in the ceramic material. In our study, for the 3 types of ceramics, the bond strength in the silica-coated group exceeded this value and showed cohesive failure in the ceramic. Although bond strength values did not exceed this value for the sandblasted and silanated samples, there were more cohesive failures than in the silica-coated group. Also, in the leuci-based group, all surface treatment groups except sandblasting showed cohesive failures in the ceramic.
In addition to shear bond testing, the surface characteristics of the porcelain after various surface-modification procedures should be examined. When the adhesive contacts the solid ceramic surface, it wets and spreads over the surface and penetrates into the pits of the roughened ceramic surface. Thus, the wetting of the roughened ceramic surface by the adhesive resin is a critical requirement for optimal bonding. Alteration of the surface topography—eg, etching—will result in changes on the surface area and on the wettability and the surface energy of the substrate. Chemical etching changed the morphologic surface of the leucite-based and lithia disilicate-based ceramics, creating an irregular topography ( Fig 4 , B and C ). HF acid can remove the glass matrix and the second crystalline phase, thus creating irregularities in the ceramic crystals. Our SEM study showed that 9.6% HF acid etching for 2 minutes is effective in the removal of the crystalline phase and the glass matrix and thus creates an retentive surface. When the feldspathic ceramic surface was treated with HF acid, uniformly distributed pores and shallow irregularities were clearly observed ( Fig 4 , A ). The chemical etching process can be explained by the preferential reaction of the HF acid with the silica phase of feldspathic ceramic. These silicates are removed by rinsing with water, and the final result is an irregular surface, suitable for bonding.
In this study, airborne particle abrasion was performed with 50-μm aluminum trioxide particles for 3 seconds. For the feldspathic, leucite-based, and lithia disilicate-based ceramics, this treatment changed their surfaces because the irregularities appeared shallower than the on the silica-coated surface ( Fig 2 ). Silica coating without silane application promoted irregularities in the 3 types of porcelain samples, but the edges of the pits and grooves were sharper than in samples treated with airborne particle abrasion ( Fig 3 ). These features might have contributed to the bond strength values of the silica-coated samples more than to the sandblasted samples.
This in-vitro study had limitations and could not fully reproduce clinical conditions. Further clinical trials are needed to obtain experience with the silica-coating technique.