INTRODUCTION
Glass ceramics are widely used to manufacture esthetic veneers, inlays, onlays, and crowns 1 . However, they are brittle and often become cracked or chipped due to secondary caries, trauma, parafunctional habits, manufacturing flaws, or stress concentration induced by occlusal adjustments 2-5 . Several factors such as cost, time, wear of sound tooth structure, and risk to pulp vitality must be considered before replacing a defective ceramic restoration 6 . Removing restorations luted with adhesive inevitably enlarges the new preparation and weakens the tooth 7 , 8 . However, certain cases can be repaired with direct composite, which is a minimally invasive, low-cost, less time-consuming procedure 2 , 5 .
The advantages of restorative repair have been routinely included for more than 10 years in most European and North American dental school syllabuses 19-11 . Some longitudinal clinical trials indicate that repaired composite restorations can remain clinically acceptable for up to 12 years 12-15 . Despite the lack of long-term evidence, the repair of ceramic restorations has shown a success rate of 89% and a survival rate of 3 years, which makes the approach feasible in certain cases 16 .
Several factors, including ceramic type, composite type, aging condition, and surface treatment protocol can influence the composite repair bond strength to ceramic restorations 6 . The success of adhesion depends on the roughness of the surface to which the composite is bonded 21 . Different protocols for intraoral repair of chipped and/or fractured ceramic restorations have been suggested to increase the bond strength to the composite: roughening by diamond burs 22 , etching with hydrofluoric acid (HF) 23 , 24 sandblasting with aluminum oxide (Al2O3) microparticles 15 , laser irradiation, and tribochemical silica coating 25 . HF etching has been widely reported as a reliable extraoral surface treatment for glass-ceramic restorations prior to adhesive luting 26 . However, in an intraoral repair scenario, HF is highly toxic and may cause severe damage to oral tissues, and its use is forbidden in dental clinics in several countries 27 . Although lower bond strength values have been reported, sandblasting with Al2O3 microparticles at adequate pressure does not harm soft tissues nor decrease the flexural strength of lithium disilicate glass-ceramics 28 .
Though most studies investigate the effect of microscopic changes promoted by different surface treatments on the ceramic surface to be repaired, it would be also relevant to address the question of whether the macroscopic design of the ceramic restoration could affect adhesive bonding to the composite repair. The aim of this study was therefore to investigate the interaction effect of different designs and surface treatments on the load to failure of lithium disilicate glass-ceramic repaired with nanofilled composite. The null hypothesis was that different macroscopic ceramic designs did not influence the load-to-failure of composite repairs, regardless of different previous surface treatments.
MATERIALS AND METHOD
Ceramic slabs 4 mm thick (10 x 10 mm) with three different designs of the top surface (flat, single 2-mm-deep plateau, or double 2-mm-deep plateau) were prototyped virtually using computer aided-design software, milled in wax using a computer aided-manufacturing unit, invested, and then heat-pressed with lithium disilicate glass-ceramic ingots (IPS e.max Press, Ivoclar Vivadent, Liechtenstein). The injection sprues were removed with diamond burs (#881 and #881F, Jota do Brasil, Florianópolis, SC, Brazil) mounted on a high-speed water-cooled air turbine, and the ceramic slabs were cleaned ultrasonically in distilled water for 30 sec ( Fig. 1 ).
The ceramic slabs of each design were assigned to subgroups (n=11) according to the following top surface treatments:
No treatment.
Etching with 5% HF (Condac Porcelain, FGM, Joinville, SC, Brazil) for 20 sec, followed by water rinsing and air-drying. One layer of silane (Relyx Ceramic Primer, 3M ESPE, Saint Paul MN, USA) was applied and air-dried after 60 sec. Then, one layer of a universal one-step adhesive (Single Bond Universal, 3M ESPE, Saint Paul, MN, USA) was applied for 20 sec, air-dried for 5 sec, and light-cured for 10 sec using an LED unit with an output of 1000 mW/cm2 (VALO, Ultradent, South Jordan, UT, USA).
Sandblasting with 50-^m Al2O3 particles for 10 sec from a distance of 5 mm. Then, both silane and adhesive were applied as described above. Each slab was placed in a polyvinyl chloride mold and incrementally repaired with nanofilled composite (Filtek Z350, 3M ESPE) up to 6 mm above the highest ceramic top plateau. The molds had marks every 2 mm to guide the thickness of each composite layer ( Fig. 2 ), which was light-cured for 20 sec on each side using the abovementioned LED unit. All specimens were stored in artificial saliva at 37 °C for 21 days (ECB 1.3 bacteriological oven, Odontobrás, Ribeirao Preto, SP, Brazil). Then, they were subjected to 1,000 thermocycles between 5 and 55 °C (30 sec dwell time) and stored in distilled water at 37 °C before testing.
The upper and the lower edges of each specimen were attached to a tensile test setup ( Fig. 3 ) and the interface composite-ceramic was tested until failure in a universal testing machine with a 200-kgf load cell (DL2000, EMIC, Sao José dos Pinhais, PR, Brazil) at a crosshead speed of 1 mm/min. The load-to-failure of each specimen was recorded in Newtons (N).
The mode of failure was determined under a stereomicroscope with 40x magnification (EK3ST, Eikonal, Sao Paulo, SP, Brazil) and classified as ‘adhesive’ (at the interface between ceramic and composite), ‘cohesive in composite’, ‘cohesive in ceramic’, or ‘mixed’ (combination of interfacial failure and cohesive in composite). One representative tested specimen of each subgroup (design/surface treatment) was sputter-coated with gold and the ceramic surface morphology was observed through scanning electron microscopy (SEM; VEGA3, Tescan, Brno, Czech Republic) at magnifications of 100X, 160X, and 500X.
Since the data did not meet the assumptions of normality and homoscedasticity, the ceramic design and surface treatment effects were analyzed using the Kruskal-Wallis test followed by Dunn’s multiple comparisons test. The failure modes were compared using G-tests. The data were analyzed with statistical software at a significance level of p<0.05 (SPSS 23.0, IBM Corp., Chicago, IL, USA; BioEstat 5.0, Mamirauá Institute, Belém, PA, Brazil).
RESULTS
Considering the flat ceramic design, no significant differences in load-to-failure values were observed between HF-etched and sandblasted specimens (Tables 1 and 2). However, HF etching resulted in significantly higher load-to-failure values than sandblasting for both single plateau and double plateau designs (Tables 3 and 4).
Regardless of ceramic design, the absence of surface treatment resulted in significantly lower load-to-failure values between lithium disilicate glass-ceramic and composite repair ( Table 1 ).
Ceramic design | No treatment | HF etching | Sandblasting | p-value |
---|---|---|---|---|
Flat | 2.19 (±4.62); 0.00 Ba | 148.62 (±66.06); 144.81 Aa | 61.29 (±50.96); 33.74 Aa | < 0.001 |
Single plateau | 5.49 (±10.99); 0.00 Ba | 171.02 (±84.40); 208.64 Aa | 16.84 (±12.58); 16.98 Bb | < 0.001 |
Double plateau | 8.90 (±17.78); 0.0 Ba | 127.59 (±64.10); 116.39 Aa | 34.34 (±35.17); 19.53 Bab | < 0.001 |
p-value | 0.733 | 0.550 | 0.031 | |
Comparisons | Rank difference | calculated Z | critical Z | p<0.05 |
---|---|---|---|---|
No treatment vs. HF etching | 18.0 | 4.5720 | 2.394 | Yes |
No treatment vs. Sandblasting | 10.2 | 2.5908 | 2.394 | Yes |
HF vs. Sandblasting | 7.8 | 1.9812 | 2.394 | No |
Comparisons | Rank difference | calculated Z | critical Z | p<0.05 |
---|---|---|---|---|
No treatment vs. HF etching | 17.95 | 4.5593 | 2.394 | Yes |
No treatment vs. Sandblasting | 5.90 | 1.4986 | 2.394 | No |
HF vs. Sandblasting | 3.0607 | 2.394 | 2.394 | Yes |
Comparisons | Rank difference | calculated Z | critical Z | p<0.05 |
---|---|---|---|---|
No treatment vs. HF etching | 18.00 | 4.5720 | 2.394 | Yes |
No treatment vs. Sandblasting | 8.10 | 2.0574 | 2.394 | No |
HF vs. Sandblasting | 9.90 | 2.5146 | 2.394 | Yes |
Approximately 80%, 70%, and 60% of the untreated specimens of flat, single plateau, and double plateau ceramic designs, respectively, had pre-testing failures during thermocycling, and their respective load-to-failure values were recorded as zero. Only 20% of sandblasted specimens with flat and single plateau designs presented pre-testing failures. Conversely, HF-etched specimens did not present failures during thermocycling.
The load-to-failure values of HF-etched specimens did not differ significantly according to the ceramic design. For specimens with sandblasted surface treatment, the load-to-failure values of flat specimens were significantly higher than for specimens with a single plateau. The double plateau design resulted in intermediate load-to-failure values, which did not differ significantly from the other two ceramic designs ( Table 5 ).
Comparisons | Rank difference | calculated Z | critical Z | p<0.05 |
---|---|---|---|---|
Flat vs. Single plateau | 10.25 | 2.6035 | 2.394 | Yes |
Flat vs. Double plateau | 6.4 | 1.6256 | 2.394 | No |
Single plateau vs. Double plateau | 3.85 | 0.9779 | 2.394 | No |
Significant differences were found among failure modes (p<0.001) Regardless of the design, untreated and sandblasted lithium disilicate glass-ceramic specimens, respectively, presented only adhesive failures and cohesive failures in composite. The percentage of cohesive failures in composite was high (80%) for HF-etched flat specimens, while most (60%) of the HF-etched specimens with single plateau or double plateau presented mixed failures. Cohesive failure in ceramic was observed only in HF-etched with single plateau design (10%) ( Fig. 4 ). SEM photomicrographs showed that the surface was smoother in HF-etched specimens than in specimens sandblasted with Al2O3 microparticles ( Fig. 5 ).
DISCUSSION
Since repairing defective ceramic restorations with direct composite can be a valuable approach due to its reliability, low cost, and conservative characteristics 29 , this study addressed the effect of macroscopic design and surface treatment on the load-to-failure of lithium disilicate glass-ceramic repaired with nanofilled composite. The null hypothesis was rejected because there was no significant difference between macroscopic ceramic design and load-to-failure values.
The bonding effectiveness of composite to ceramic depends strongly on micromechanical retention 6 , so the lithium disilicate glass-ceramic surface was roughened by HF etching or Al2O3 sandblasting. Regardless of the ceramic design, the highest values of load-to-failure were observed for specimens etched with 5% HF before composite repair. Although the microstructure of lithium disilicate glass-ceramic has high crystal content, HF etching dissolves the glassy matrix of ceramic, creating a superficial porous microretentive surface which increases the surface free energy and wettability for adhesive bonding 30 , 31 . Both HF concentration and etching time were within the acceptable range that does not jeopardize the bond strength to lithium disilicate glass-ceramic 15 , 24 .
The results also demonstratedthatAl2O3 sandblasting created a certain amount of micromechanical retention on lithium disilicate glass-ceramic. Flat ceramic surfaces resulted in significantly higher load-to-failure values than did single plateau design, suggesting that Al2O3 sandblasting is less effective when applied on angulated ceramic walls. Moreover, sandblasting was significantly less effective than HF etching for both single plateau and double plateau ceramic designs; however, the difference was not significant between HF-etched and sandblasted flat ceramic surfaces. Sandblasting increases the surface roughness and surface area of glass-ceramics; however, surface roughness above certain microlevels can form microcracks that reduce micromechanical retention and decrease bond strength; in addition, deep irregular pits on the ceramic surface do not provide retentional features 32 . The application of a silane coupling agent produces a chemical link between the silicate in the ceramic surface and the polymer-based hydrophobic components in the composite through covalent siloxane bonds 4 , 33 . Thus, the combination of mechanical and chemical retention increases the bond strength of ceramic and repair composite 34 . The results of the current study corroborate some other studies that recommend HF etching followed by silanization as the gold standard surface treatment for silica-based glass-ceramics 4,5,35The specimens were stored in artificial saliva for 21 days and thermocycled between 5 and 55 oC (1,000 cycles) because these techniques are widely accepted to simulate aging of the interface between ceramic and composite repair 6 , 8 . Regardless of the ceramic design, 80% of the specimens that did not receive surface treatment presented pre-testing failure during thermocycling, which indicates the importance of creating microretention on the ceramic surface before composite repair.
In the HF-etched specimens, the load-to-failure values were higher for single plateau ceramic design than for flat design. Although the double plateau was expected to provide even more retention for the composite repair, it had the lowest load-to-failure values. Since the load-to-failure values of all ceramic designs were relatively high and did not differ significantly, it seems that the failures (mostly cohesive in composite or mixed) occurred due to intrinsic characteristics of the composite. Furthermore, the relatively high standard deviations presented by most of the groups may be related to the macro design of the specimens, in which the occurrence of internal gaps along the ceramic/ composite interface as well as stress accumulation, particularly at the internal corners of single and double plateau specimens, may have influenced the overall results 36 .
All lithium disilicate glass-ceramic specimens that did not receive previous surface treatment failed at the interface between ceramic and composite, which was not observed in any HF-etched or sandblasted specimen. Regardless of the ceramic design, all sandblasted specimens presented cohesive failure in composite, which suggests that the interfacial bond strength provided by sandblasting is higher than the cohesive strength of the nanofilled composite. Moreover, the higher load to failure values of specimens with flat design in comparison to both single plateau and double plateau indicates that a thick, uniform layer of repair composite when sandblasting is used as ceramic surface treatment.
In contrast to flat specimens, HF-etched specimens with single or double plateau presented the most mixed failures, which suggests better stress distribution throughout the composite and the lithium disilicate glass-ceramic. The optical profilometry analysis conducted by Lima et al. (2021) showed that both sandblasting with 50-pm Al2O3 particles and silica coating with 30-pm Al2O3 particles resulted in the most pronounced alterations on the ceramic surface. The authors reported that although sandblasting created the highest surface roughness, it also promoted surface damage in all evaluated ceramic types. Moreover, 10% HF etching increased flexural strength, particularly when applied for 20 sec.
Strasser et al. (2018) reported that HF provided strong, homogenous etching patterns on lithium disilicate glass-ceramic, in which the glass phase was dissolved and the crystals were found to be relatively protruded. Nevertheless, sandblasting with 50-pm Al2O3 particles resulted in the highest values of surface roughness. Gul & Uygun (2020) reported that sandblasting caused the most remarkable alterations on ceramic surfaces. In the current study, the SEM images of sandblasted specimens also showed deeper, more irregular roughness than HF-etched specimens. In addition, the Al2O3 microparticles potentially abraded the lithium disilicate glass matrix and crystals to a certain level that weakened the surface.
The results of this study therefore indicated that the surface treatment of a defective lithium disilicate glass-ceramic restoration has more influence on the retention of the composite repair than its macroscopic design. Moreover, the highest load-to-failure values, the failure mode pattern, and the regular surface roughness observed for HF-etched lithium disilicate glass-ceramic specimens suggest higher bond strength to the composite repair.