Metal-free materials used as dental restorations have been in the spotlight of recent research, following the introduction of innovative all-ceramic materials (1). Also, the rising demand for more facial esthetics has increased the request for adult orthodontics (2). However, orthodontic brackets bond poorly to ceramic surfaces, unless the surface characteristics of the ceramic are altered through certain approaches before bonding (3).
Earlier studies reported that bond strength of brackets to various restorations can be connected to many factors, such as restoration material and its surface conditioning, the material and the design of the bracket, the properties of the bond system, as well as the light-curing device. Hereupon, the combination of these and other factors may be very important for successful treatment (4-6).
In recent past, we witnessed fast high-tech evolution of various all-ceramic dental restorations (6, 7). Partially stabilized zirconia is a very common choice in dentistry, because of its indisputable mechanical properties, chemical inertness and excellent optical properties (7). Another hot-spot all-ceramic that includes excellent esthetics with proper strength is lithium disilicate (7–9).
The type of ceramic used as restoration can be a decisive factor for the binding of orthodontic attachments and for the method of altering their surface before bonding. In the past, various surface treatment methods of the ceramic surface were introduced, such as diamond burs, sandblasting, hydrofluoric acid (HFA), phosphoric acid (PhA), laser etching, etc., exposing their advantages and disadvantages (10–15). Mechanical removal of the glazed surface of the ceramic with diamond burs as well as sandblasting aluminum-oxide particles with high pressure can enhance bond strength, but also can reduce ceramic integrity, which could lead to cracks and larger damages during debonding (16). HFA etching creates a porous surface by removing the glassy matrix (17) and has been shown to result in acceptable bond-strength values in porcelain (11, 14, 15), but it is less successful in more crystalline rich ceramics. Also, the danger of acid burns is very high, which can result in deep tissue necrosis (18). Conditioning with some lasers has also been investigated, also with not so satisfactory results (19–21). Phosphoric acid (37.0%) cannot erode superficial layers of silicate ceramic (10, 11, 13), but it is able to neutralize the alkalinity of the absorbed water present on ceramic restorations in the mouth and thereby making better chemical conditions for the subsequently applied silane (22). It is also not toxic or corrosive and in combination with silane achieves satisfactory bond strength (13, 22). The use of silane increases the adhesion of the composite resin bond to ceramics (11, 13, 23), by creating a chemical link between the hydroxyl (OH) of the silica of the ceramic with the resinous matrix of the composite (6, 17, 24). However, with the increase of the crystalline phase in the content of the ceramic, this chemical reaction becomes less efficient, because of lower levels of silica (17).
When bonding onto ceramic surfaces, the material type consisting orthodontic brackets and their base surface design should also be considered. Oftentimes, adhesion of ceramic brackets is higher compared to metallic brackets, because of the better light transmittance allowing stronger photo-polymerization. This is also due to a different failure mode because of the flexibility of the metal base (4, 6, 13, 24, 25).
Currently, there is no consensus regarding the most efficient ceramic surface conditioning method for producing optimal bond strength of orthodontic attachments to various ceramic surfaces. The purpose of this research was to conduct a comparative analysis of Shear Bond Strength (SBS) and of Adhesive Remnant Index (ARI), or ceramic and metallic brackets bonded to zirconia and lithium disilicate ceramic surfaces used for prosthetic restorations, conditioned with 5% HFA or 37% PhA, and silane. Also, the objective of this research was to investigate the Porcelain Fracture Index (PFI), in order to examine the condition of the ceramic surface after debonding.
Material and methods
Preparation of specimens: The research was conducted on 96 all-ceramic crowns, of which 48 full contour zirconia (Copran Zr-i Monolith, White Peaks Dental Solutions GmbH&Co.KG, Wesel, Essen, Germany), and 48 lithium disilicate (IPS EMAX CAD, Ivoclar Vivadent AG Schaan, Lichtenstein). Metallic orthodontic brackets (Mini 2000 Ormco Corp., Glendora, California, USA) and ceramic orthodontic brackets (Glam Forestadent, Bernhard Forster GmbH, Pforzheim, Germany) were equally bonded to these crowns. Two different etching materials were used for conditioning of the surface of ceramic crowns: 5% HFA (IPS Ceramic Etching Gel, Ivoclar Vivadent AG, Schaan, Lichtenstein) or 37% PhA (Etching solution, Ormco Corp., Glendora, CA, USA) for 120 s, and subsequently silane (Prosil, Dentscare, Joinville, Brasil) was applied. Two-component (primer and adhesive) composite resin-based bonding system (Tranbond XT, 3M/Unitek, Monrovia, CA, USA) was used for bonding the brackets. The brackets were bonded in the middle of the prepared surfaces of the ceramic sample by the same operator. They were pressed firmly, and the excess adhesive was removed from around the bracket base using a dental probe. The adhesive was light-cured for 40 s (4), using a light-emitting diode (LED; Edition, Ivoclar Vivadent AG, Schaan, Lichtenstein). Prior to testing, the crowns were embedded in a two-component epoxy filling (Epoxy Repair, Bison International, Goes, The Netherlands). Additionally, five days after bonding the brackets, the specimens were thermocycled (5800 cycles, 5ºC to 55ºC in distilled water, with 10 s dwelling time), in order to simulate the moisture of the oral environment.
Groups: The sample was divided into 8 groups: 1. Metallic bracket bonded to zirconia surface etched with PhA; 2. Metallic bracket bonded to zirconia surface etched with HFA; 3. Ceramic bracket bonded to zirconia surface etched with PhA; 4. Ceramic bracket bonded to zirconia surface etched with HFA; 5. Metallic bracket bonded to lithium disilicate surface etched with PhA; 6. Metallic bracket bonded to lithium disilicate surface etched with HFA; 7. Ceramic bracket bonded to lithium disilicate surface etched with PhA; and 8. Ceramic bracket bonded to lithium disilicate surface etched with HFA.
SBS testing: SBS was tested with Universal Testing Machine (Erichsen 0-2000 N, ISO 7500-1:1, AM Erichsen GmbH&Co.KG, Hemer-Sundwig, Germany), with a load applied parallel to the buccal surface of the crown in a gingival-occlusal direction, using a knife-edged rod moving at fixed rate of 1 mm/min, until failure occurred. The force required to debond the brackets was recorded in Newton, and the values were calculated to MPa. (Figure 1).
SEM examination and ARI determination: After debonding, the samples were analyzed using Scanning Electron Microscope (Tescan Vega TS5136MM, Chez Rep) and photomicrographs were taken, to determine Adhesive Remnant Index (ARI) and Porcelain Fracture Index (PFI). Before examination under SEM, the samples were dehydrated for 5h, in increasing concentrations of alcohol (70% and 95%). Subsequently, the non-conducting materials (ceramic brackets and both types of crowns) were coated with gold and palladium sputter (SC7620 Mini Sputter Coater, Quorum Technologies Ltd, UK). Furthermore, a localized chemical analysis was performed with Energy Dispersive X-ray Spectrometry (EDS) in representative samples, in order to determine the concentrations of the elements present in both types of ceramics.
In order to determine ARI (as per Bishara et al.) (25), the measurements were performed, using scores varying from 1 to 5: 1 - All adhesive remaining on the ceramic crown surface with the impression of the bracket base; 2 - More than 90% of the adhesive remaining on the ceramic crown surface; 3 - Less than 90%, but more than 10% of the adhesive remaining on the surface; 4 - Less than 10% of the adhesive remaining on the ceramic crown surface; and 5 - No adhesive remaining on the ceramic crown surface.
Damage to the ceramic surface which may have occurred during shear bond testing was recorded using PFI (Bourke and Rock) (13). The index was divided into four scores as follows: 0 - ceramic surface intact or in the same condition as before the bonding procedure; 1 - surface damage limited to glaze layer or very superficial ceramic; 2 - surface damage which features significant loss of ceramic requiring restoration of the defect by composite resin or replacement of the restoration; 3 - surface damage where the core material has been exposed due to the depth of the cohesive failure.
Statistical analysis: The Kolmogorov-Smirnov test was used to test the distribution of SBS data. The hypothesis that SBS is similar in all the groups was tested using the univariate analysis of variance (ANOVA) with the ceramic type, bracket type and etching method as random factors. The significance level was set at 0.05. The data were analyzed using STATISTICA 10 (StatSoft, Inc., version 10, www.statsoft.com).
The results of SBS by ceramic type, bracket type, and etching method are shown in Table 1.
SD – Standard deviation
According to the results of the Kolmogorov-Smirnov test, SBS values are normally distributed (Kolmogorov-Smirnov Z = 0.721, p = 0.676). Test for equality of variances (Hartley F-max = 4.92, Bartlett Chi-Sqr. = 7.33, Cochran C = 0.206, p = 0.396) confirms that there is no significant difference between the variables of the individual subsamples i.e. the samples are homogeneous.
The results obtained from the univariate test of significance are presented in Table 2. They indicate that none of the factors or their interaction have a significant influence on the mean values of SBS. According to that, and explanations based on the Cohen criteria, variations of SBS are minor and in this case not significant.
The results are presented in Figure 2, where the mean SBS values and corresponding confidence interval (0.95) are shown for the 8 groups generated from all three factors. Significant overlap is noticeable in these intervals.
According to the results of the chi-square test, the frequency of certain categories of ARI significantly depends only on the type of the bracket (chi-square = 14.85, df = 4, p = 0.005). The first and the second category of ARI appear to have a significantly higher frequency regarding metallic brackets: first ARI category occurs only with metallic brackets, and the second in 87.5% of the cases also occurs with metallic brackets. The frequency of the third category of ARI is equally distributed in both types of brackets. Again, the fourth category of ARI is more frequent in ceramic brackets with 63.6%. In the fifth category of ARI, the frequency is almost equal with a slight advantage of ceramic brackets (53.8%). This is also illustrated in Figure 3 and also SEM photomicrographs are presented in Figure 4.
Porcelain fracture index (PFI) significantly differs from the type of etchant (chi-square = 4.746, df = 1, p = 0.029). The first PFI category (0) occurs significantly more frequently with the preparation of the substrate with PhA, while the other category (1) is more frequently present with the preparation of substrates with HFA (Table 3). The two last categories were noticed in neither of the examined samples.
a count, b % within PFI
The SEM photomicrographs of the two ceramic surfaces etched with HFA revealed different surface morphologies. Zirconia ceramic displayed fewer pits and more unchanged glazed surfaces than the lithium disilicate ceramic. In both type of ceramic, the crowns etched with PhA, loss of the glazed surface and mild roughening were observed. Uniform peeling or an erosive appearance with shallow penetration and undercuts was also observed (Figure 5).
Noticeable differences in chemical element concentrations of measurements with EDS between lithium disilicate (42% O, 32% Si, 9.3% K, 8.6% C, 3% Al, 2.6% W, 2.1% Zn) and zirconia (48.6% Zr, 43.2% O, 5.7% Si, 1% Al, 0.9 K) are presented in Figure 6.
When bonding brackets to ceramic surfaces, double challenges arise. In order to avoid bond failure, the optimal bond strength of 6 to 10 MPa during the treatment is recommended (13). Again, after debonding, the restorations should remain in the same condition with their ideal esthetic and function (4, 26). Nevertheless, transferring this value in clinical work is questionable because of the complex environment of the oral cavity (14). In this research, the mean SBS values for all combinations were more than 6 MPa, but less than 13 MPa, which may cause cracks in the ceramic (26).
According to numerous studies (4, 6, 13, 24, 25), the bond strength of ceramic brackets is higher than the strength of metallic brackets. But, our results indicate that this doesn’t occur at orthodontic brackets bonded to zirconia restorations, and this is in accordance with a study previously reported by Mehmeti et al. (2017) (27), where metallic brackets, in comparison with ceramic brackets, bond better with zirconia restorations. This might be because of the base surface design of metallic brackets, producing a better mechanical coupling with zirconia ceramic substrate. Furthermore, this was not the case in most of our groups with ceramic brackets, except group 7 showing the best result, however still not significant in comparison with metallic bracket groups. In general, ceramic brackets bonded to lithium disilicate samples, compared to those bonded to zirconia, showed slightly but not significantly higher SBS values. The highest difference between the lithium disilicate and zirconia was registered in a ceramic bracket and phosphoric acid groups, probably due to the variations in the chemical compositions of these two ceramic materials.
Regardless of the type of ceramic and its surface conditioning, the samples with metallic brackets have shown mixed adhesive-cohesive failures. In the majority of the samples with ceramic brackets, adhesive failures were noticed (scale 4 or 5), which indicates that the bond strength between the composite and the ceramic bracket was stronger than the bond strength between the composite and ceramic crown. This type of failure is desired to avoid ceramic breakage during debonding (28). Our findings are partially in concordance with the above mentioned findings.
Furthermore, in neither of the all-ceramic types larger fractures or cracks were observed, which is clinically important for the long-term integrity of the restoration. The significant difference between two etchants regarding PFI that was noticed is in agreement with other studies (6, 13) and may indicate that the use of HFA can make more vulnerable the ceramic surface of both all-ceramic materials.
Our findings may indicate that the use of HFA is unnecessary for conditioning the ceramic surface before bonding orthodontic brackets. However, this research was conducted under in vitro conditions, which are not always possible to compare with clinical situations. According to Bourk and Rock (1999), thermocycling weakens bond strength and is recommended in order to simulate the conditions in the oral cavity (13). On the other hand, Smith et al. (1988) stated that thermocycling had no significant effect on SBS (28). However, in this research, thermocycling was performed as a mean of artificial aging of the bond prior to testing.
Despite limitations, SBS testing remains a relevant methodology to compare bonding protocols by providing important information regarding bracket debonding in clinical situations (29).
The results of this paper lead us to the conclusion that the use of HFA for surface etching of zirconia and/or lithium disilicate does not cause a significant rise of the SBS values in comparison to etch with PhA and silane application. Furthermore, HFA can weaken the surface structure of the ceramic, and considering its adverse effect, might not be the best suitable conditioner prior to orthodontic bonding to lithium disilicate, and in particular to zirconia, also taking into account its crystalline structure.
Regarding the orthodontic point of view, zirconia and lithium disilicate all-ceramic restorations, as well as both types of brackets and both type of etchants, have similar features and provide strong enough values to ensure appropriate treatment.