- Email: [email protected]
Evaluation of microtensile bond strength on ceramic-resin adhesion using two specimen testing substrates
International Journal of Adhesion & Adhesives 54 (2014) 165–171
Contents lists available at ScienceDirect
International Journal of Adhesion & Adhesives journal homepage: www.elsevier.com/locate/ijadhadh
Evaluation of microtensile bond strength on ceramic-resin adhesion using two specimen testing substrates Tian Tian a, James Kit-Hon Tsoi b, Jukka P. Matinlinna b, Michael F. Burrow a,n a b
Oral Diagnosis and Polyclinics, Faculty of Dentistry, University of Hong Kong, Hong Kong Special Administrative Region, PR China Dental Materials Science, Faculty of Dentistry, University of Hong Kong, Hong Kong Special Administrative Region, PR China
art ic l e i nf o
a b s t r a c t
Article history: Accepted 16 May 2014 Available online 12 June 2014
Objectives: The objective of this study was to compare two bonding models using a microtensile bond test and evaluate the effect of two surface treatments on lithium disilicate ceramics using two resin cements. Methods: Ceramic blocks (e.max CADs) were sectioned, polished and fired for final crystallization. The blocks were treated with one of two surface treatments: (1) hydrofluoric acid (HF) (IPS Ceramic Etching gel) etched followed by silane (Monobond-S) application; (2) HF etched, silane applied, followed by hot air drying and rinsed with hot water, dried and an unfilled resin (Heliobond) applied. Ceramics without surface treatment were the control. Two bonding substrates were used: resin composite and ceramic with the same surface treatment and the corresponding groups were divided into two bonding models: ceramic to ceramic (C–C) and ceramic to resin composite (C–R). Two resin cements, Variolink IIs and Clearfil SA Cement, were tested. Each group (n ¼ 30) was stored in distilled water for 7 days at 37 1C, then subjected to a tensile force until failure. Failure modes were determined with stereomicroscope and SEM. ANOVA, Bonferroni tests and Weibull analysis were used for statistical analysis (po 0.05). Results: All the control groups experienced spontaneous debonding during preparation. The C–C groups showed significantly higher bond strength than the C–R groups (po 0.05). Failure mode in the C–R groups was dominated by cohesive failure in resin cement while in the C–C groups was mostly mixed failure. Ceramic treated with HF etching and silanization and luted with Variolink II showed the highest bond strength (53.5 7 6.6 MPa) while ceramic treated with HF etching, silanization and hot treatment and luted with Clearfil SA Cement showed the lowest bond strength (35.477.0 MPa) in the C–C groups. Weibull analysis showed that Weibull modulus in the C–C model was higher than the C–R model. Conclusions: Ceramic-bonded to ceramic model is recommended for evaluating the microtensile bond strength of ceramic-resin cement-adhesion. Variolink II showed better bonding than Clearfil SA Cement. & 2014 Elsevier Ltd. All rights reserved.
Keywords: Glass ceramics Resin bonding Microtensile bond strength Surface modification Weibull analysis
1. Introduction Nowadays glass ceramics are widely used in dental restorations because of their excellent esthetic performance. However, a major limitation for clinical use of these materials is their comparatively low strength which may lead to fracture and restoration failure [6,7]. IPS e.max Press (Ivoclar Vivadent AG, Schaan, Liechtenstein), a lithium disilicate ceramic, has been introduced with improved mechanical properties compared with other glass ceramics [1]. Recently this system was further enhanced and has exhibited a higher strength and better processing technology than IPS e.max n Correspondence to: Oral Diagnosis and Polyclinics, Faculty of Dentistry, University of Hong Kong, Prince Philip Dental Hospital, 34 Hospital Road, Sai Ying Pun, Hong Kong, PR China. E-mail address: [email protected] (M.F. Burrow).
http://dx.doi.org/10.1016/j.ijadhadh.2014.06.003 0143-7496/& 2014 Elsevier Ltd. All rights reserved.
Press. This newer ceramic, IPS e.max CAD, can now be clinically used for 3- to 4-unit bridges and for constructing milled restorations. The bond to ceramics is of great importance to the long-term success of glass ceramic restorations. Nowadays, maximal preservation of dental hard tissue is advocated in dental treatment [23]. There are increasing cases where the retention of restorations is chiefly reliant on adhesion with the tooth preparation having minimal traditional retentive form, e.g. short crown height. Hence, a durable and reliable bond is required for success of such ceramic restorations. Compared with traditional ‘adhesive’ luting cements such as zinc polycarboxylate or glass ionomer cement, resin composite luting cements have been introduced to overcome retention problems of all-ceramic restorations. They not only provide a stronger and more durable bond between ceramic materials and tooth structure, but are also more esthetic. Furthermore, it has been found that the ceramic strength is enhanced by the use of resin cements.
166
T. Tian et al. / International Journal of Adhesion & Adhesives 54 (2014) 165–171
The adhesion of ceramic restorations has two interfaces: the ceramic-resin cement interface and resin cement-tooth interface. A number of studies have been performed evaluating the ceramicresin bond with the aim of increasing the adhesive strength [10,18,20]. It is well recognized that bonding between ceramics and resin cements can be achieved by two mechanisms: micromechanical attachment and a chemical bond between the ceramic and resin cement. Micromechanical adhesion is created by HF etching [12] and/or grit blasting while a chemical bond is achieved with a silane coupling agent [14]. However, no wide agreement has been reached for the optimal bonding methodology. The traditional method of surface treatment is HF etching followed by silane application to the fitting surface of the ceramic. Nevertheless, a study has shown various bonding procedures to a leucite reinforced ceramic showed an ‘optimal’ method which was a particular silane treatment with heating of the silanated ceramic [11]. Due to the different chemical composition and microstructure between leucite reinforced and lithium disilicate ceramics, this method may not be the best choice for lithium disilicate ceramics. So far, no study has been performed on testing the bond between lithium disilicate ceramics and resin cement using this ‘optimal’ method mentioned above. Of the various laboratory bond test methods available, the microtensile bond strength test is a relatively reliable method which has the potential to reveal a “true” bond strength as failure is of the adhesive. In most studies, to date, the test method employed for resin-ceramic bonding evaluation has used a bonding model that consists of a ceramic block bonded to a resin composite block with a resin luting cement [16,17]. In this model, the bond strength that is of interest is at the interface of the resin cement and ceramic on account that the adhesion to the resin composite is regarded as similar to the resin cement, as well as resin cement to composite is not often used clinically. However, there are cases where fracture may occur at the interface between the resin cement and resin composite or cohesively in the resin composite material. This form of failure is not at the adhesive interface of interest, therefore if such data are included in a study, it will lead to inaccurate outcomes of the test aim. An alternative model is a combination of two ceramic blocks ‘luted’ with a resin cement [11]. In this model, the fracture would be generated either at the interface of resin-ceramic or cohesively in the resin cement. The results are believed to be more indicative of the ‘true’ adhesive strength of bonded interface. However, no study has been carried out evaluating the difference of these two bonding models. Therefore, the aim of this study was to compare the two bonding models (ceramic to resin composite and ceramic to ceramic) using a microtensile bond strength test. In addition, an evaluation of the effect of surface treatments between the traditional method and the ‘optimal’ method used in studies mentioned above was undertaken using two resin cements. The null hypotheses tested were: (1) there is no difference in microtensile bond strength between two bonding substrates; and (2) there is no difference in microtensile bond strength among different surface treatments and resin luting cements.
2. Materials and methods The materials, batch number and manufacturers are listed in Table 1. 2.1. Preparation of ceramics Ceramic blocks (e.max CADs, Table 1) were sectioned into smaller blocks (14 12 3 mm3 thick) with a self-assembled cutting machine (Miki pulley DC motor, Miki Pulley Co., Ltd, Japan)
using an alloy blade. Each block was further polished with 180 , 400 , 600 and 1200-grit SiC papers using a polishing machine (ECOMET 5, Buehler, Düsseldorf, Germany) under running water for 30 min with light hand pressure. The polished surfaces were then ultrasonically cleansed in 95% ethanol for 5 min and air dried for 5 min. After that, the blocks were fired for final crystallization using Programat CS ceramic furnace (Ivoclar Vivadent AG, Schaan, Liechtenstein) following the manufacturer's instructions for e. max CAD. The polished surfaces were treated using two different surface treatment methods. Specimens in the first treatment group were etched with 4.7% Hydrofluoric acid (Table 1) for 60 s and rinsed with running tap water for 60 s and ultrasonically cleaned with distilled water for 5 min. A silane coupling agent (Table 1) was applied to the etched surface for 60 s. This is the conventional surface treatment method for glass ceramics. Specimens in the other treatment group were etched following the method used in the first group, then the ‘optimal’ method of silane treatment was applied as described in a previous study [11]. The ‘optimal’ method is, the silane was applied for 60 s, then hot air dried at 50 75 1C for 15 s with a hair dryer, then the dried silanated surfaces were dipped into 80 1C distilled water for 15 s, dried with hot air for 30 s, then a thin layer of unfilled resin (Table 1) was applied to the treated surface. Polished surfaces without any surface treatment were used as control groups. 2.2. Preparation of resin composite Resin composite (Table 1) blocks were made 14 12 3 mm3 thick, the same dimensions as the ceramic block, by layering 3 mm-thick increments in a silicone mold, and light-cured for 40 s for each increment with a LED light curing unit (EliparTM 2500, 3 M ESPE, Seefeld, Germany). The last increment was condensed using a flat glass slide to obtain a flat surface. The five surfaces of the composite that contacted with the silicone mold were further irradiated for another 40 s for each side after the removal of the specimens (total 200 s). The blocks were then taken to be heat cured at 100 1C for 15 min following the manufacturer's instructions. The to-be-bonded surface was polished with 500-grit SiC paper under running water followed by ultrasonic cleansing in distilled water for 5 min and air dried for 5 min. 2.3. Luting procedure Two bonding models were used in this experiment. The first was two ceramic blocks ‘luted’ with one of the two resin cements, while the other model used a ceramic block ‘luted’ to a resin composite block with resin cement. The specimens were further subdivided into two groups by using two different resin cements: Variolink IIs (Table 1) and Clearfil SA Cement (Table 1). The bonding procedures were carried out using a loading device that produced a constant seating load of 2 N (0.2 kg), applied for 5 min when the two blocks were bonded together. Excess cement was removed with a scalpel blade before complete setting of the cement. Specimens were photo-activated for 20 s on each side of the block with the LED curing light (total 80 s). The ceramic-tocomposite luting procedures were repeated in the same manner. 2.4. Microtensile bond strength test Each bonded block was fixed with sticky wax to the platform of the cutting machine, and was vertically sectioned into a series of 1 mm thick slices using a water-cooled rotating diamond saw blade. The block was rotated 901 and the cutting procedure repeated. Beams were obtained from each block with approximately 1.0 mm2 cross-sectional areas. Thirty beams were prepared
T. Tian et al. / International Journal of Adhesion & Adhesives 54 (2014) 165–171
167
Table 1 Overview of the experimental materials. Material
Brand
Manufacturer
Composition (Information supplied by the manufacturer)
Batch no.
Glass ceramics
e.max CADs
SiO2, Li2O, K2O, P2O5, ZrO2, ZnO, Al2O3, MgO, Coloring oxides
R28555
Resin cement
Variolink IIs
bis-GMA, TEGDMA, UDMA, inorganic fillers (Ytterbium trifluoride, barium glass, Ba-Al-fluorosilicate glass), catalysts and stabilizers, pigments
Base: R35481 Catalyst: P84939
Resin cement
Clearfil SA Cement
Ivoclar Vivadent AG, Schaan, Liechtenstein Ivoclar Vivadent AG, Schaan, Liechtenstein Kuraray Noritake Dental Inc., Tokyo, Japan
00007C
Silane coupling agent
Monobond-S
MDP, bis-GMA, TEGDMA, hydrophobic aromatic dimethacrylate, hydrophobic aliphatic dimethacrylate, silanated barium glass filler, Silanated colloidal silica, dl-Camphorquinone, Benzoyl peroxide, surface treated sodium fluoride, initiator, accelerators Ethanol, 3-trimethoxysilylpropyl methacrylate
P70737
HF acid
IPS Ceramic etching gel
Hydrofluoric acid
M47179
Adhesive
Heliobond
bis-GMA, TEGDMA, initiators, stabilizers
R23145
Resin composite
ESTENIA™ C&B
Polyurethane methacrylmonomer, methacrylic acid series monomer, Filler (Surface treated glass powder and surface treated aluminum micro filler), photocuring catalyst, colorant and others
0094AB
Ivoclar Vivadent AG, Schaan, Liechtenstein Ivoclar Vivadent AG, Schaan, Liechtenstein Ivoclar Vivadent AG, Schaan, Liechtenstein Kuraray Noritake Dental Inc., Tokyo, Japan
TEGDMA: triethylenglycol dimethacrylate, bis-GMA: bisphenol-A-diglycidyl methacrylate, UDMA: urethane dimethacrylate, MDP: 10-methacryloyloxydecyl dihydrogen phosphate.
for each group and were stored in distilled water at 37 1C for 7 days prior to the bond test. After storage, the cross sectional area of each stick was measured with digital calipers (Mitutoyo, Tokyo, Japan). Specimens were attached using a cyanoacrylate glue (Zapit CA, Dental Ventures of America, Corona, California, USA) to the two ends of the mirotensile test jig [9] mounted on a universal loading machine (Instron 4444, Norwood, Massachusetts, USA). A tensile stress was applied at a crosshead speed of 0.5 mm/min until bond rupture occurred. The load at failure was recorded and the microtensile bond strength (σ) value was calculated using the equation σ ¼L/A, where L was the failure load and A the crosssectional area. The failure modes of each specimen were determined by examining both sides of the fractured specimens using a stereomicroscope (100 ) and scanning electron microscope (SEM) (Hitachi S-3400N VP-SEM, Hitachi High-Technologies, Tokyo, Japan). In this study, the failure modes were classified as cohesive failure in the resin cement; interfacial failure/adhesive failure; mixed failure, and spontaneous failure [21]. A cohesive failure was defined as failure occurring in either the resin cement without exposure of the ceramic/composite surface or the adherends (ceramic and resin composite). Interfacial failure/adhesive failure was designated as failure at the adhesive/ceramic interface. Mixed failure represented a combination of adhesive and cohesive failure. Pre-testing failure (PTF) was for those the samples that fractured after the beams were prepared or prior to mounting on the test jig. SEM observations were also performed on the ceramic surfaces treated with different surface treatments and Energy-dispersive X-ray (EDX) analysis was done to investigate the elemental composition of the two resin cements.
2.5. Statistical analysis Data were statistically analyzed using SPSS for Windows (Version 20.0, SPSS Inc., Chicago, IL, USA). Three-way ANOVA and one-way ANOVA with further Bonferroni multiple comparison were used at 95% significance level. The microtensile bond strength data were further analyzed using Weibull analysis
software package (R and weibulltoolkit 2.2, R Foundation for Statistical Computing, Vienna, Austria) [13].
3. Results 3.1. Microtensile bond strength Control groups bonded with Variolink II or Clearfil SA Cement in both bonding models all experienced spontaneous debonding during the preparation of specimens. These four groups were excluded from the statistical analysis. The mean microtensile bond strength data for each group are shown in Table 2 and Fig. 1. Significant differences were revealed between the two surface treatments and two bond test specimen types (p o0.05). Three-way ANOVA was carried out to analyze the effect of bonding substrate on the bond strength. The result showed that ceramic to ceramic model had a significantly higher bond strength than ceramic to resin composite model. Weibull parameters, including the Weibull modulus (β), the 63.2% of the samples expected to fail (η) and the 10.0% of the samples expected to fail (B10), are presented in Table 2 and Weibull plots are shown in Fig. 2. Weibull modulus in each ceramic to ceramic group was higher compared to the corresponding group in the ceramic to resin groups. The level at which 63.2% of samples would be expected to fail, as well as the 10.0% level, showed overall higher values in the ceramic to ceramic groups compared with the ceramic to resin groups. 3.2. Failure modes, SEM observation and EDX analysis Results of failure modes are shown in Table 3. In the ceramic bonded to resin composite groups, the dominant failure mode was cohesive failure in the resin cement (Fig. 3(C)) (Table 3). For the ceramic bonded to ceramic groups, most groups were dominated by mixed failure, with the occurrence of mixed failures (Fig. 3 (D) and (E)) and adhesive failures (Fig. 3(A)) being greater than for the ceramic bonded to resin composite groups. The failure modes of the control groups bonded with Clearfil SA Cement or Variolink II in both bonding models are not presented in Table 3.
168
T. Tian et al. / International Journal of Adhesion & Adhesives 54 (2014) 165–171
Table 2 Mean microtensile bond strength and Weibull parameters with different surface treatments, bonding models and resin cements (n¼ 30). Group C–R C–R C–R C–R C–C C–C C–C C–C
Clearfil SA Cement Clearfil SA Cement Variolink II (1) Variolink II (2) Clearfil SA Cement Clearfil SA Cement Variolink II (1) Variolink II (2)
Mean strength (SD) (MPa) (1) (2)
(1) (2)
31.6 30.1 24.8 18.9 36.3 35.4 53.5 41.1
(4.8)a (5.8)a (7.7)b (6.0)c (6.7)ac (7.0)a (6.6)b (10.0)c
βa
ηb
7.8 6.4 4.4 3.6 6.8 6.6 9.9 5.1
33.5 32.2 27.0 20.9 38.8 37.8 56.2 44.5
B10c (30.7–36.5) (29.1–35.9) (23.3–31.8) (17.4–25.2) (35.1–43.2) (34.2–41.9) (52.6–60.2) (39.3–50.7)
25.1 22.6 16.1 11.2 27.8 26.9 44.7 28.7
(19.2–29.1) (16.5–27.1) (9.8–21.0) (6.0–15.4) (20.1–33.0) (19.2–32.1) (36.1–50.3) (19.0–36.0)
C–R represents ceramic bonded to resin composite groups; C–C represent ceramic bonded to ceramic groups; (1) indicates ceramic surface was treated with HF acid for 60 s, rinsed, silane coupling agent applied for 60 s; (2) indicates ceramic surface was treated with HF acid for 60 s, rinsed, silane coupling agent applied for 60 s, hot air for 15 s, hot water rinsed for 15 s, hot air for 30 s and unfilled resin applied; Groups with different superscript letters represent significant differences in the same bonding substrate groups (p o 0.05). a b c
Represents β or modulus Weibull parameter. Represents η, estimation and 90% confidence interval at 63.2% of the investigated populations expected to fail. Represents estimation of strength and 90% confidence interval at which 10% of the investigated populations would be expected to fail.
Fig. 1. Microtensile bond strength with different surface treatments, bonding models and resin cements (median, lower/1st and upper/3rd quartiles and maxima/minima). C–R represents ceramic bonded to resin composite groups; C–C represent ceramic bonded to ceramic groups; (1) indicates ceramic surface was treated with HF acid for 60 s, rinsed, silane coupling agent applied for 60 s; (2) indicates ceramic surface was treated with HF acid for 60 s, rinsed, silane coupling agent applied for 60 s, hot air for 15 s, hot water rinsed for 15 s, hot air for 30 s and unfilled resin applied.
SEM micrographs of the different ceramic surface treatments are shown in Fig. 4. A porous surface can be seen after the HF etching of the ceramic surface (Fig. 4(B)). Elements of the two resin cements identified by EDX are demonstrated in Fig. 5. EDX measurements showed that Clearfil SA Cement contains oxygen, silicon, aluminum, barium, phosphorus, and sodium (Fig. 5(A)). Variolink II presented oxygen, fluoride, phosphorus, calcium, ytterbium, silicon and barium (Fig. 5(B)).
4. Discussion In this study, the μTBS was analyzed using Weibull statistics. Weibull analysis has been developed as a method that is routinely used in structural engineering to determine the characteristic failure of brittle materials. Since many materials used in restorative dentistry are brittle, this methodology is gaining popularity in dental research, especially for ceramic strength [22] and bond strength testing [4,19]. The advantage of this analytical method is that multiple parameters are provided to describe the behavior of a material as well as being able to predict and forecast when failure of a material may occur at a particular stress level, rather
than simply evaluate the mean failure strength of each material. Two typical parameters are presented in every analysis. One is β, the slope or the shape of the Weibull curve, also called the Weibull modulus. It is a measurement of the variability or distribution of the data. Another parameter is η, the scale parameter. This is called the characteristic life or characteristic strength in the case of this study and represents the estimated value at which 63.2% of the investigated population is expected to fail. The value of 10% probability of failure is also presented in this study because it may reflect when early failures may occur in the clinical situation. A limited number of studies are available in the dental literature regarding Weibull analysis of bond strength of resin bonded to ceramics [3,5,8,15]. However, all data presented were from older types of glass ceramics. The Weibull modulus obtained in this study (3.6–9.9) is in the similar range of previously recorded moduli (2.1–14.1). Only one study has previously measured the microtensile bond strength of a lithium disilicate ceramic bonded to resin composite [5]. When the ceramic surface was treated with the same surface treatment (HF etching and silanization), the Weibull modulus in our study showed a relatively lower value compared with previously published value as well as the characteristic strength (η). The possible reason might
T. Tian et al. / International Journal of Adhesion & Adhesives 54 (2014) 165–171
169
Fig. 2. Weibull plots for the microtensile bond strength tests. C–R represents ceramic bonded to resin composite groups; C–C represents ceramic bonded to ceramic groups; (1) indicates ceramic surface was treated with HF acid for 60 s, rinsed, silane coupling agent applied for 60 s; (2) indicates ceramic surface was treated with HF acid for 60 s, rinsed, silane coupling agent applied for 60 s, hot air for 15 s, hot water rinsed for 15 s, hot air for 30 s and unfilled resin applied.
Table 3 Failure modes of each group. Group
Adhesive failure
Mixed failure
Cohesive Cohesive failure failure in resin cement in resin composite
C–R Clearfil SA Cement (1) C–R Clearfil SA Cement (2) C–R Variolink II C–R Variolink II C–C Clearfil SA Cement (1) C–C Clearfil SA Cement (2) C–C Variolink II C–C Variolink II
1/30
12/30
15/30
2/30
0/30
4/30
25/30
1/30
0/30 0/30 1/30
1/30 0/30 20/30
27/30 30/30 9/30
2/30 0/30 /
6/30
19/30
5/30
/
3/30 0/30
8/30 18/30
19/30 12/30
/ /
(1) (2)
(1) (2)
be that resin cement was not used in the previous study. The previous study used a resin composite filling material that was applied directly after the adhesive application. This might create fewer flaws and thus leading to a higher Weibull modulus or be an effect of the difference in elastic modulus between resin cement and resin composite. A further aspect is the variation of stress distribution across the adhesive interface between the two bonding models. It is possible this difference in stress distribution also influenced the outcomes. This is a complex and broad topic which is beyond the scope of the current work and requires a separate study to elucidate potential factors that could influence the bond strength and failure pattern. When comparing the two bonding substrates (resin composite and ceramic), a general trend observed was that the ceramic bond was usually higher to ceramic than the bond to the resin composite substrate, as shown by the mean μTBS and Weibull parameters in each ceramic to ceramic group which were higher compared to the corresponding group in the ceramic to resin groups (Table 2). The overall higher Weibull modulus in the ceramic to ceramic groups indicates that this model might create
more reliable data and thus lead to a more ‘valid’ result. Also, the failure mode of the μTBS specimens revealed that more specimens failed as mixed and/or adhesive failure in the ceramic to ceramic groups (Table 3). These two failure modes can be considered to be closer to the “true” failure at the interface. However, it was noted that the mixed failure in the ceramic to ceramic groups were more complicated, that is, the crack tended to propagate across the two opposing interfaces in various directions rather than simply failing at one interface only. Whether this failure pattern can be regarded as adhesive failure or not is open to debate and further study is necessary. Since the ceramic bonded to ceramic model yielded the advantage of seeming more reliable, it may reveal a closer value to the “true” bond strength at the interface of the ceramic and resin cement. Clinically, however, the bonding procedure of ceramic restorations is bonding a ceramic restoration to tooth structure (dentine or enamel) or resin composite core materials. Crack propagation and failure mode of the restoration might be different from the ceramic to ceramic model. This model might not be reproduced in actual clinical practice. Consequently, the data of the ceramic to ceramic model should be treated carefully by not directly interpreting into a clinical scenario because it may not reflect the true crack propagation and failure mode in a restoration bonded to tooth structure. The ceramic to resin composite model, on the contrary, is occasionally seen in a restoration when a core of the tooth is built up by resin composite and bonded to a ceramic restoration. In addition, resin composite is used as a dentine substitute in laboratory studies because it is generally accepted that the mechanical properties of these two materials are similar. However, the bond strength should still be interpreted carefully. This is because the data obtained in such a model revealed not only the bonding potential at the interface of ceramic and resin cement, but also included the mechanical properties of resin composite and resin cements on the basis of failure modes (cohesive failure in resin composite/cement). Since the ceramic to ceramic model is relatively more reliable in revealing the bond potential at the interface of ceramic and resin cement, a comparison between Clearfil SA Cement and Variolink II is consequently more reasonably explained using this
170
T. Tian et al. / International Journal of Adhesion & Adhesives 54 (2014) 165–171
Fig. 3. Examples of typical SEM images of failure modes (100 ) (A) Adhesive failure; (B) Cohesive failure in resin composite; (C) Cohesive failure in resin cement; (D) Mixed failure (C–C model); (E) Mixed failure (C–R model).
Fig. 4. Ceramic surface with or without HF etching (5000 ) (A) Ceramic surface polishing without HF etching; (B) Polished ceramic surface followed by HF etching.
Fig. 5. Elements identified by EDX for Clearfil SA Cement (A) and Variolink II (B).
model. Variolink II showed significantly higher bonding potential to glass ceramics than Clearfil SA Cement (Table 2 and Fig. 1). Although the elements from the two resin cements are similar by
showing high concentration in silicon and barium (Fig. 5), the mean diameter of the filler particles in Variolink II is smaller than that in Clearfil SA Cement, 0.7 μm and 2.5 μm respectively. Resin
T. Tian et al. / International Journal of Adhesion & Adhesives 54 (2014) 165–171
cement with smaller filler particles may be adapting more intimately to the roughened surface and thus result in improved bonding. Another possible reason is the silane that was used in this study. Variolink II exhibited better bonding performance when used with the silane (Monobond S), possibly as these two materials are produced from the same manufacturer and thus highly compatible. A recent study also showed that Variolink II performed better than Clearfil SA Cement when using Monobond S [2]. However, it is unclear if Variolink II or Clearfil SA Cement would achieve the same performance if a different silane coupling agent was applied. Further study is needed. The conventional surface treatment on ceramics showed better bonding behavior in both bonding models than the one recommended in the previous study mentioned (Table 2). This is quite apparent in the groups using Variolink II resin cement. The possible reason might be related to the different ceramic structures used in this study compared with the previous work. In the previous study, a modified leucite reinforced ceramic was with silane application without HF etching. In the present study, however, a lithium disilicate ceramic was used with HF etching followed by silanization. This may indicate that micromechanical attachment created by HF etching is still important in ceramic surface preparation to achieve good adhesion. The first hypothesis that there is no difference in microtensile bond strength between two bonding substrates was rejected. The second hypothesis that there is no difference in microtensile bond strength among different surface treatments and resin luting cements was also rejected.
5. Conclusions Comparing the two bonding models, it is clear that ceramic substrate yields some advantages over the resin composite substrate by providing more reliable and consistent results. However, considering its complex situation in mixed failure, the clinical relevance should be more carefully interpreted. This model is probably more useful for evaluating adhesion of resin materials to ceramics. Consequently, it is necessary to further the improvements on the methodology of ceramic bonding evaluation.
Acknowledgment The authors are grateful to the following companies for providing the experimental materials used in this study: Ivoclar Vivadent, Kuraray Noritake Dental and Dental Ventures of America.
171
References [1] Albakry M, Guazzato M, Swain MV. Biaxial flexural strength, elastic moduli, and X-ray diffraction characterization of three pressable all-ceramic materials. J Prosthet Dent 2003;89:374–80. [2] Bahr N, Keul C, Edelhoff D, Eichberger M, Roos M, Gernet W, et al. Effect of different adhesives combined with two resin composite cements on shear bond strength to polymeric CAD/CAM materials. Dent Mater J 2013;32:492–501. [3] Boscato N, Della Bona A, Del Bel Cury AA. Influence of ceramic pre-treatments on tensile bond strength and mode of failure of resin bonded to ceramics. Am J Dent 2007;20:103–8. [4] Burrow MF, Thomas D, Swain MV, Tyas MJ. Analysis of tensile bond strengths using Weibull statistics. Biomaterials 2004;25:5031–5. [5] Della Bona A, Anusavice KJ, Mecholsky Jr. JJ. Failure analysis of resin composite bonded to ceramic. Dent Mater: Off Publ Acad Dent Mater 2003;19:693–9. [6] Frankenberger R, Taschner M, Garcia-Godoy F, Petschelt A, Kramer N. Leucitereinforced glass ceramic inlays and onlays after 12 years. J Adhes Dent 2008;10:393–8. [7] Hayashi M, Tsuchitani Y, Kawamura Y, Miura M, Takeshige F, Ebisu S. Eightyear clinical evaluation of fired ceramic inlays. Oper Dent 2000;25:473–81. [8] Hernandez AI, Roongruangphol T, Katsube N, Seghi RR. Residual interface tensile strength of ceramic bonded to dentin after cyclic loading and aging. J Prosthet Dent 2008;99:209–17. [9] Ho G, Matinlinna JP. Evaluation of the microtensile bond strength between resin composite and hydrofluoric acid etched ceramic in different storage media. J Adhes Sci Technol 2011;25:2671–85. [10] Hooshmand T, Rostami G, Behroozibakhsh M, Fatemi M, Keshvad A, van Noort R. Interfacial fracture toughness of different resin cements bonded to a lithium disilicate glass ceramic. J Dent 2012;40:139–45. [11] Hooshmand T, van Noort R, Keshvad A. Bond durability of the resin-bonded and silane treated ceramic surface. Dent Mater: Off Publi Acad Dent Mater 2002;18:179–88. [12] Horn HR. Porcelain laminate veneers bonded to etched enamel. Dent Clin North Am 1983;27:671–84. [13] J Symynck, F.d.B., Monte Carlo pivotal confidence bounds for Weibull analysis, with implementations in R; May 13–14 2011. [14] Matsumura H, Kato H, Atsuta M. Shear bond strength to feldspathic porcelain of two luting cements in combination with three surface treatments. J. Prosthet. Dent. 1997;78:511–7. [15] Michelini FS, Belser UC, Scherrer SS, De Rijk WG. Tensile bond strength of gold and porcelain inlays to extracted teeth using three cements. Int J Prosthodont 1995;8:324–31. [16] Pelogia F, Valandro LF, Brigagao V, Neisser MP, Bottino MA. Resin microtensile bond strength to feldspathic ceramic: hydrofluoric acid etching vs. tribochemical silica coating. Int J Prosthodont 2007;20:532–4. [17] Pisani-Proenca J, Erhardt MC, Valandro LF, Gutierrez-Aceves G, BolanosCarmona MV, Del Castillo-Salmeron R, et al. Influence of ceramic surface conditioning and resin cements on microtensile bond strength to a glass ceramic. J Prosthet Dent 2006;96:412–7. [18] Pollington S, Fabianelli A, van Noort R. Microtensile bond strength of a resin cement to a novel fluorcanasite glass-ceramic following different surface treatments. Dent Mater: Off Publ Acad Dent Mater 2010;26:864–72. [19] Roos M, Stawarczyk B. Evaluation of bond strength of resin cements using different general-purpose statistical software packages for two-parameter Weibull statistics. Dent Mater: Off Publ Acad Dent Mater 2012;28:e76–88. [20] Sakai M, Taira Y, Sawase T. Silane primers rather than heat treatment contribute to adhesive bonding between tri-n-butylborane resin and a machinable leucite-reinforced ceramic. Dent mater J 2011. [21] Soderholm KJM. Critical evaluation of adhesive test methods used in dentistry. J Adhes Sci Technol 2009;23:973–90. [22] Stawarczyk B, Ozcan M, Trottmann A, Hammerle CH, Roos M. Evaluation of flexural strength of hipped and presintered zirconia using different estimation methods of Weibull statistics. J Mech Behavi Biomed Mater 2012;10:227–34. [23] Tyas MJ, Anusavice KJ, Frencken JE, Mount GJ. Minimal intervention dentistry —a review. FDI Commission Project 1–97. Int. Dent. J. 2000;50:1–12.
Contents lists available at ScienceDirect
International Journal of Adhesion & Adhesives journal homepage: www.elsevier.com/locate/ijadhadh
Evaluation of microtensile bond strength on ceramic-resin adhesion using two specimen testing substrates Tian Tian a, James Kit-Hon Tsoi b, Jukka P. Matinlinna b, Michael F. Burrow a,n a b
Oral Diagnosis and Polyclinics, Faculty of Dentistry, University of Hong Kong, Hong Kong Special Administrative Region, PR China Dental Materials Science, Faculty of Dentistry, University of Hong Kong, Hong Kong Special Administrative Region, PR China
art ic l e i nf o
a b s t r a c t
Article history: Accepted 16 May 2014 Available online 12 June 2014
Objectives: The objective of this study was to compare two bonding models using a microtensile bond test and evaluate the effect of two surface treatments on lithium disilicate ceramics using two resin cements. Methods: Ceramic blocks (e.max CADs) were sectioned, polished and fired for final crystallization. The blocks were treated with one of two surface treatments: (1) hydrofluoric acid (HF) (IPS Ceramic Etching gel) etched followed by silane (Monobond-S) application; (2) HF etched, silane applied, followed by hot air drying and rinsed with hot water, dried and an unfilled resin (Heliobond) applied. Ceramics without surface treatment were the control. Two bonding substrates were used: resin composite and ceramic with the same surface treatment and the corresponding groups were divided into two bonding models: ceramic to ceramic (C–C) and ceramic to resin composite (C–R). Two resin cements, Variolink IIs and Clearfil SA Cement, were tested. Each group (n ¼ 30) was stored in distilled water for 7 days at 37 1C, then subjected to a tensile force until failure. Failure modes were determined with stereomicroscope and SEM. ANOVA, Bonferroni tests and Weibull analysis were used for statistical analysis (po 0.05). Results: All the control groups experienced spontaneous debonding during preparation. The C–C groups showed significantly higher bond strength than the C–R groups (po 0.05). Failure mode in the C–R groups was dominated by cohesive failure in resin cement while in the C–C groups was mostly mixed failure. Ceramic treated with HF etching and silanization and luted with Variolink II showed the highest bond strength (53.5 7 6.6 MPa) while ceramic treated with HF etching, silanization and hot treatment and luted with Clearfil SA Cement showed the lowest bond strength (35.477.0 MPa) in the C–C groups. Weibull analysis showed that Weibull modulus in the C–C model was higher than the C–R model. Conclusions: Ceramic-bonded to ceramic model is recommended for evaluating the microtensile bond strength of ceramic-resin cement-adhesion. Variolink II showed better bonding than Clearfil SA Cement. & 2014 Elsevier Ltd. All rights reserved.
Keywords: Glass ceramics Resin bonding Microtensile bond strength Surface modification Weibull analysis
1. Introduction Nowadays glass ceramics are widely used in dental restorations because of their excellent esthetic performance. However, a major limitation for clinical use of these materials is their comparatively low strength which may lead to fracture and restoration failure [6,7]. IPS e.max Press (Ivoclar Vivadent AG, Schaan, Liechtenstein), a lithium disilicate ceramic, has been introduced with improved mechanical properties compared with other glass ceramics [1]. Recently this system was further enhanced and has exhibited a higher strength and better processing technology than IPS e.max n Correspondence to: Oral Diagnosis and Polyclinics, Faculty of Dentistry, University of Hong Kong, Prince Philip Dental Hospital, 34 Hospital Road, Sai Ying Pun, Hong Kong, PR China. E-mail address: [email protected] (M.F. Burrow).
http://dx.doi.org/10.1016/j.ijadhadh.2014.06.003 0143-7496/& 2014 Elsevier Ltd. All rights reserved.
Press. This newer ceramic, IPS e.max CAD, can now be clinically used for 3- to 4-unit bridges and for constructing milled restorations. The bond to ceramics is of great importance to the long-term success of glass ceramic restorations. Nowadays, maximal preservation of dental hard tissue is advocated in dental treatment [23]. There are increasing cases where the retention of restorations is chiefly reliant on adhesion with the tooth preparation having minimal traditional retentive form, e.g. short crown height. Hence, a durable and reliable bond is required for success of such ceramic restorations. Compared with traditional ‘adhesive’ luting cements such as zinc polycarboxylate or glass ionomer cement, resin composite luting cements have been introduced to overcome retention problems of all-ceramic restorations. They not only provide a stronger and more durable bond between ceramic materials and tooth structure, but are also more esthetic. Furthermore, it has been found that the ceramic strength is enhanced by the use of resin cements.
166
T. Tian et al. / International Journal of Adhesion & Adhesives 54 (2014) 165–171
The adhesion of ceramic restorations has two interfaces: the ceramic-resin cement interface and resin cement-tooth interface. A number of studies have been performed evaluating the ceramicresin bond with the aim of increasing the adhesive strength [10,18,20]. It is well recognized that bonding between ceramics and resin cements can be achieved by two mechanisms: micromechanical attachment and a chemical bond between the ceramic and resin cement. Micromechanical adhesion is created by HF etching [12] and/or grit blasting while a chemical bond is achieved with a silane coupling agent [14]. However, no wide agreement has been reached for the optimal bonding methodology. The traditional method of surface treatment is HF etching followed by silane application to the fitting surface of the ceramic. Nevertheless, a study has shown various bonding procedures to a leucite reinforced ceramic showed an ‘optimal’ method which was a particular silane treatment with heating of the silanated ceramic [11]. Due to the different chemical composition and microstructure between leucite reinforced and lithium disilicate ceramics, this method may not be the best choice for lithium disilicate ceramics. So far, no study has been performed on testing the bond between lithium disilicate ceramics and resin cement using this ‘optimal’ method mentioned above. Of the various laboratory bond test methods available, the microtensile bond strength test is a relatively reliable method which has the potential to reveal a “true” bond strength as failure is of the adhesive. In most studies, to date, the test method employed for resin-ceramic bonding evaluation has used a bonding model that consists of a ceramic block bonded to a resin composite block with a resin luting cement [16,17]. In this model, the bond strength that is of interest is at the interface of the resin cement and ceramic on account that the adhesion to the resin composite is regarded as similar to the resin cement, as well as resin cement to composite is not often used clinically. However, there are cases where fracture may occur at the interface between the resin cement and resin composite or cohesively in the resin composite material. This form of failure is not at the adhesive interface of interest, therefore if such data are included in a study, it will lead to inaccurate outcomes of the test aim. An alternative model is a combination of two ceramic blocks ‘luted’ with a resin cement [11]. In this model, the fracture would be generated either at the interface of resin-ceramic or cohesively in the resin cement. The results are believed to be more indicative of the ‘true’ adhesive strength of bonded interface. However, no study has been carried out evaluating the difference of these two bonding models. Therefore, the aim of this study was to compare the two bonding models (ceramic to resin composite and ceramic to ceramic) using a microtensile bond strength test. In addition, an evaluation of the effect of surface treatments between the traditional method and the ‘optimal’ method used in studies mentioned above was undertaken using two resin cements. The null hypotheses tested were: (1) there is no difference in microtensile bond strength between two bonding substrates; and (2) there is no difference in microtensile bond strength among different surface treatments and resin luting cements.
2. Materials and methods The materials, batch number and manufacturers are listed in Table 1. 2.1. Preparation of ceramics Ceramic blocks (e.max CADs, Table 1) were sectioned into smaller blocks (14 12 3 mm3 thick) with a self-assembled cutting machine (Miki pulley DC motor, Miki Pulley Co., Ltd, Japan)
using an alloy blade. Each block was further polished with 180 , 400 , 600 and 1200-grit SiC papers using a polishing machine (ECOMET 5, Buehler, Düsseldorf, Germany) under running water for 30 min with light hand pressure. The polished surfaces were then ultrasonically cleansed in 95% ethanol for 5 min and air dried for 5 min. After that, the blocks were fired for final crystallization using Programat CS ceramic furnace (Ivoclar Vivadent AG, Schaan, Liechtenstein) following the manufacturer's instructions for e. max CAD. The polished surfaces were treated using two different surface treatment methods. Specimens in the first treatment group were etched with 4.7% Hydrofluoric acid (Table 1) for 60 s and rinsed with running tap water for 60 s and ultrasonically cleaned with distilled water for 5 min. A silane coupling agent (Table 1) was applied to the etched surface for 60 s. This is the conventional surface treatment method for glass ceramics. Specimens in the other treatment group were etched following the method used in the first group, then the ‘optimal’ method of silane treatment was applied as described in a previous study [11]. The ‘optimal’ method is, the silane was applied for 60 s, then hot air dried at 50 75 1C for 15 s with a hair dryer, then the dried silanated surfaces were dipped into 80 1C distilled water for 15 s, dried with hot air for 30 s, then a thin layer of unfilled resin (Table 1) was applied to the treated surface. Polished surfaces without any surface treatment were used as control groups. 2.2. Preparation of resin composite Resin composite (Table 1) blocks were made 14 12 3 mm3 thick, the same dimensions as the ceramic block, by layering 3 mm-thick increments in a silicone mold, and light-cured for 40 s for each increment with a LED light curing unit (EliparTM 2500, 3 M ESPE, Seefeld, Germany). The last increment was condensed using a flat glass slide to obtain a flat surface. The five surfaces of the composite that contacted with the silicone mold were further irradiated for another 40 s for each side after the removal of the specimens (total 200 s). The blocks were then taken to be heat cured at 100 1C for 15 min following the manufacturer's instructions. The to-be-bonded surface was polished with 500-grit SiC paper under running water followed by ultrasonic cleansing in distilled water for 5 min and air dried for 5 min. 2.3. Luting procedure Two bonding models were used in this experiment. The first was two ceramic blocks ‘luted’ with one of the two resin cements, while the other model used a ceramic block ‘luted’ to a resin composite block with resin cement. The specimens were further subdivided into two groups by using two different resin cements: Variolink IIs (Table 1) and Clearfil SA Cement (Table 1). The bonding procedures were carried out using a loading device that produced a constant seating load of 2 N (0.2 kg), applied for 5 min when the two blocks were bonded together. Excess cement was removed with a scalpel blade before complete setting of the cement. Specimens were photo-activated for 20 s on each side of the block with the LED curing light (total 80 s). The ceramic-tocomposite luting procedures were repeated in the same manner. 2.4. Microtensile bond strength test Each bonded block was fixed with sticky wax to the platform of the cutting machine, and was vertically sectioned into a series of 1 mm thick slices using a water-cooled rotating diamond saw blade. The block was rotated 901 and the cutting procedure repeated. Beams were obtained from each block with approximately 1.0 mm2 cross-sectional areas. Thirty beams were prepared
T. Tian et al. / International Journal of Adhesion & Adhesives 54 (2014) 165–171
167
Table 1 Overview of the experimental materials. Material
Brand
Manufacturer
Composition (Information supplied by the manufacturer)
Batch no.
Glass ceramics
e.max CADs
SiO2, Li2O, K2O, P2O5, ZrO2, ZnO, Al2O3, MgO, Coloring oxides
R28555
Resin cement
Variolink IIs
bis-GMA, TEGDMA, UDMA, inorganic fillers (Ytterbium trifluoride, barium glass, Ba-Al-fluorosilicate glass), catalysts and stabilizers, pigments
Base: R35481 Catalyst: P84939
Resin cement
Clearfil SA Cement
Ivoclar Vivadent AG, Schaan, Liechtenstein Ivoclar Vivadent AG, Schaan, Liechtenstein Kuraray Noritake Dental Inc., Tokyo, Japan
00007C
Silane coupling agent
Monobond-S
MDP, bis-GMA, TEGDMA, hydrophobic aromatic dimethacrylate, hydrophobic aliphatic dimethacrylate, silanated barium glass filler, Silanated colloidal silica, dl-Camphorquinone, Benzoyl peroxide, surface treated sodium fluoride, initiator, accelerators Ethanol, 3-trimethoxysilylpropyl methacrylate
P70737
HF acid
IPS Ceramic etching gel
Hydrofluoric acid
M47179
Adhesive
Heliobond
bis-GMA, TEGDMA, initiators, stabilizers
R23145
Resin composite
ESTENIA™ C&B
Polyurethane methacrylmonomer, methacrylic acid series monomer, Filler (Surface treated glass powder and surface treated aluminum micro filler), photocuring catalyst, colorant and others
0094AB
Ivoclar Vivadent AG, Schaan, Liechtenstein Ivoclar Vivadent AG, Schaan, Liechtenstein Ivoclar Vivadent AG, Schaan, Liechtenstein Kuraray Noritake Dental Inc., Tokyo, Japan
TEGDMA: triethylenglycol dimethacrylate, bis-GMA: bisphenol-A-diglycidyl methacrylate, UDMA: urethane dimethacrylate, MDP: 10-methacryloyloxydecyl dihydrogen phosphate.
for each group and were stored in distilled water at 37 1C for 7 days prior to the bond test. After storage, the cross sectional area of each stick was measured with digital calipers (Mitutoyo, Tokyo, Japan). Specimens were attached using a cyanoacrylate glue (Zapit CA, Dental Ventures of America, Corona, California, USA) to the two ends of the mirotensile test jig [9] mounted on a universal loading machine (Instron 4444, Norwood, Massachusetts, USA). A tensile stress was applied at a crosshead speed of 0.5 mm/min until bond rupture occurred. The load at failure was recorded and the microtensile bond strength (σ) value was calculated using the equation σ ¼L/A, where L was the failure load and A the crosssectional area. The failure modes of each specimen were determined by examining both sides of the fractured specimens using a stereomicroscope (100 ) and scanning electron microscope (SEM) (Hitachi S-3400N VP-SEM, Hitachi High-Technologies, Tokyo, Japan). In this study, the failure modes were classified as cohesive failure in the resin cement; interfacial failure/adhesive failure; mixed failure, and spontaneous failure [21]. A cohesive failure was defined as failure occurring in either the resin cement without exposure of the ceramic/composite surface or the adherends (ceramic and resin composite). Interfacial failure/adhesive failure was designated as failure at the adhesive/ceramic interface. Mixed failure represented a combination of adhesive and cohesive failure. Pre-testing failure (PTF) was for those the samples that fractured after the beams were prepared or prior to mounting on the test jig. SEM observations were also performed on the ceramic surfaces treated with different surface treatments and Energy-dispersive X-ray (EDX) analysis was done to investigate the elemental composition of the two resin cements.
2.5. Statistical analysis Data were statistically analyzed using SPSS for Windows (Version 20.0, SPSS Inc., Chicago, IL, USA). Three-way ANOVA and one-way ANOVA with further Bonferroni multiple comparison were used at 95% significance level. The microtensile bond strength data were further analyzed using Weibull analysis
software package (R and weibulltoolkit 2.2, R Foundation for Statistical Computing, Vienna, Austria) [13].
3. Results 3.1. Microtensile bond strength Control groups bonded with Variolink II or Clearfil SA Cement in both bonding models all experienced spontaneous debonding during the preparation of specimens. These four groups were excluded from the statistical analysis. The mean microtensile bond strength data for each group are shown in Table 2 and Fig. 1. Significant differences were revealed between the two surface treatments and two bond test specimen types (p o0.05). Three-way ANOVA was carried out to analyze the effect of bonding substrate on the bond strength. The result showed that ceramic to ceramic model had a significantly higher bond strength than ceramic to resin composite model. Weibull parameters, including the Weibull modulus (β), the 63.2% of the samples expected to fail (η) and the 10.0% of the samples expected to fail (B10), are presented in Table 2 and Weibull plots are shown in Fig. 2. Weibull modulus in each ceramic to ceramic group was higher compared to the corresponding group in the ceramic to resin groups. The level at which 63.2% of samples would be expected to fail, as well as the 10.0% level, showed overall higher values in the ceramic to ceramic groups compared with the ceramic to resin groups. 3.2. Failure modes, SEM observation and EDX analysis Results of failure modes are shown in Table 3. In the ceramic bonded to resin composite groups, the dominant failure mode was cohesive failure in the resin cement (Fig. 3(C)) (Table 3). For the ceramic bonded to ceramic groups, most groups were dominated by mixed failure, with the occurrence of mixed failures (Fig. 3 (D) and (E)) and adhesive failures (Fig. 3(A)) being greater than for the ceramic bonded to resin composite groups. The failure modes of the control groups bonded with Clearfil SA Cement or Variolink II in both bonding models are not presented in Table 3.
168
T. Tian et al. / International Journal of Adhesion & Adhesives 54 (2014) 165–171
Table 2 Mean microtensile bond strength and Weibull parameters with different surface treatments, bonding models and resin cements (n¼ 30). Group C–R C–R C–R C–R C–C C–C C–C C–C
Clearfil SA Cement Clearfil SA Cement Variolink II (1) Variolink II (2) Clearfil SA Cement Clearfil SA Cement Variolink II (1) Variolink II (2)
Mean strength (SD) (MPa) (1) (2)
(1) (2)
31.6 30.1 24.8 18.9 36.3 35.4 53.5 41.1
(4.8)a (5.8)a (7.7)b (6.0)c (6.7)ac (7.0)a (6.6)b (10.0)c
βa
ηb
7.8 6.4 4.4 3.6 6.8 6.6 9.9 5.1
33.5 32.2 27.0 20.9 38.8 37.8 56.2 44.5
B10c (30.7–36.5) (29.1–35.9) (23.3–31.8) (17.4–25.2) (35.1–43.2) (34.2–41.9) (52.6–60.2) (39.3–50.7)
25.1 22.6 16.1 11.2 27.8 26.9 44.7 28.7
(19.2–29.1) (16.5–27.1) (9.8–21.0) (6.0–15.4) (20.1–33.0) (19.2–32.1) (36.1–50.3) (19.0–36.0)
C–R represents ceramic bonded to resin composite groups; C–C represent ceramic bonded to ceramic groups; (1) indicates ceramic surface was treated with HF acid for 60 s, rinsed, silane coupling agent applied for 60 s; (2) indicates ceramic surface was treated with HF acid for 60 s, rinsed, silane coupling agent applied for 60 s, hot air for 15 s, hot water rinsed for 15 s, hot air for 30 s and unfilled resin applied; Groups with different superscript letters represent significant differences in the same bonding substrate groups (p o 0.05). a b c
Represents β or modulus Weibull parameter. Represents η, estimation and 90% confidence interval at 63.2% of the investigated populations expected to fail. Represents estimation of strength and 90% confidence interval at which 10% of the investigated populations would be expected to fail.
Fig. 1. Microtensile bond strength with different surface treatments, bonding models and resin cements (median, lower/1st and upper/3rd quartiles and maxima/minima). C–R represents ceramic bonded to resin composite groups; C–C represent ceramic bonded to ceramic groups; (1) indicates ceramic surface was treated with HF acid for 60 s, rinsed, silane coupling agent applied for 60 s; (2) indicates ceramic surface was treated with HF acid for 60 s, rinsed, silane coupling agent applied for 60 s, hot air for 15 s, hot water rinsed for 15 s, hot air for 30 s and unfilled resin applied.
SEM micrographs of the different ceramic surface treatments are shown in Fig. 4. A porous surface can be seen after the HF etching of the ceramic surface (Fig. 4(B)). Elements of the two resin cements identified by EDX are demonstrated in Fig. 5. EDX measurements showed that Clearfil SA Cement contains oxygen, silicon, aluminum, barium, phosphorus, and sodium (Fig. 5(A)). Variolink II presented oxygen, fluoride, phosphorus, calcium, ytterbium, silicon and barium (Fig. 5(B)).
4. Discussion In this study, the μTBS was analyzed using Weibull statistics. Weibull analysis has been developed as a method that is routinely used in structural engineering to determine the characteristic failure of brittle materials. Since many materials used in restorative dentistry are brittle, this methodology is gaining popularity in dental research, especially for ceramic strength [22] and bond strength testing [4,19]. The advantage of this analytical method is that multiple parameters are provided to describe the behavior of a material as well as being able to predict and forecast when failure of a material may occur at a particular stress level, rather
than simply evaluate the mean failure strength of each material. Two typical parameters are presented in every analysis. One is β, the slope or the shape of the Weibull curve, also called the Weibull modulus. It is a measurement of the variability or distribution of the data. Another parameter is η, the scale parameter. This is called the characteristic life or characteristic strength in the case of this study and represents the estimated value at which 63.2% of the investigated population is expected to fail. The value of 10% probability of failure is also presented in this study because it may reflect when early failures may occur in the clinical situation. A limited number of studies are available in the dental literature regarding Weibull analysis of bond strength of resin bonded to ceramics [3,5,8,15]. However, all data presented were from older types of glass ceramics. The Weibull modulus obtained in this study (3.6–9.9) is in the similar range of previously recorded moduli (2.1–14.1). Only one study has previously measured the microtensile bond strength of a lithium disilicate ceramic bonded to resin composite [5]. When the ceramic surface was treated with the same surface treatment (HF etching and silanization), the Weibull modulus in our study showed a relatively lower value compared with previously published value as well as the characteristic strength (η). The possible reason might
T. Tian et al. / International Journal of Adhesion & Adhesives 54 (2014) 165–171
169
Fig. 2. Weibull plots for the microtensile bond strength tests. C–R represents ceramic bonded to resin composite groups; C–C represents ceramic bonded to ceramic groups; (1) indicates ceramic surface was treated with HF acid for 60 s, rinsed, silane coupling agent applied for 60 s; (2) indicates ceramic surface was treated with HF acid for 60 s, rinsed, silane coupling agent applied for 60 s, hot air for 15 s, hot water rinsed for 15 s, hot air for 30 s and unfilled resin applied.
Table 3 Failure modes of each group. Group
Adhesive failure
Mixed failure
Cohesive Cohesive failure failure in resin cement in resin composite
C–R Clearfil SA Cement (1) C–R Clearfil SA Cement (2) C–R Variolink II C–R Variolink II C–C Clearfil SA Cement (1) C–C Clearfil SA Cement (2) C–C Variolink II C–C Variolink II
1/30
12/30
15/30
2/30
0/30
4/30
25/30
1/30
0/30 0/30 1/30
1/30 0/30 20/30
27/30 30/30 9/30
2/30 0/30 /
6/30
19/30
5/30
/
3/30 0/30
8/30 18/30
19/30 12/30
/ /
(1) (2)
(1) (2)
be that resin cement was not used in the previous study. The previous study used a resin composite filling material that was applied directly after the adhesive application. This might create fewer flaws and thus leading to a higher Weibull modulus or be an effect of the difference in elastic modulus between resin cement and resin composite. A further aspect is the variation of stress distribution across the adhesive interface between the two bonding models. It is possible this difference in stress distribution also influenced the outcomes. This is a complex and broad topic which is beyond the scope of the current work and requires a separate study to elucidate potential factors that could influence the bond strength and failure pattern. When comparing the two bonding substrates (resin composite and ceramic), a general trend observed was that the ceramic bond was usually higher to ceramic than the bond to the resin composite substrate, as shown by the mean μTBS and Weibull parameters in each ceramic to ceramic group which were higher compared to the corresponding group in the ceramic to resin groups (Table 2). The overall higher Weibull modulus in the ceramic to ceramic groups indicates that this model might create
more reliable data and thus lead to a more ‘valid’ result. Also, the failure mode of the μTBS specimens revealed that more specimens failed as mixed and/or adhesive failure in the ceramic to ceramic groups (Table 3). These two failure modes can be considered to be closer to the “true” failure at the interface. However, it was noted that the mixed failure in the ceramic to ceramic groups were more complicated, that is, the crack tended to propagate across the two opposing interfaces in various directions rather than simply failing at one interface only. Whether this failure pattern can be regarded as adhesive failure or not is open to debate and further study is necessary. Since the ceramic bonded to ceramic model yielded the advantage of seeming more reliable, it may reveal a closer value to the “true” bond strength at the interface of the ceramic and resin cement. Clinically, however, the bonding procedure of ceramic restorations is bonding a ceramic restoration to tooth structure (dentine or enamel) or resin composite core materials. Crack propagation and failure mode of the restoration might be different from the ceramic to ceramic model. This model might not be reproduced in actual clinical practice. Consequently, the data of the ceramic to ceramic model should be treated carefully by not directly interpreting into a clinical scenario because it may not reflect the true crack propagation and failure mode in a restoration bonded to tooth structure. The ceramic to resin composite model, on the contrary, is occasionally seen in a restoration when a core of the tooth is built up by resin composite and bonded to a ceramic restoration. In addition, resin composite is used as a dentine substitute in laboratory studies because it is generally accepted that the mechanical properties of these two materials are similar. However, the bond strength should still be interpreted carefully. This is because the data obtained in such a model revealed not only the bonding potential at the interface of ceramic and resin cement, but also included the mechanical properties of resin composite and resin cements on the basis of failure modes (cohesive failure in resin composite/cement). Since the ceramic to ceramic model is relatively more reliable in revealing the bond potential at the interface of ceramic and resin cement, a comparison between Clearfil SA Cement and Variolink II is consequently more reasonably explained using this
170
T. Tian et al. / International Journal of Adhesion & Adhesives 54 (2014) 165–171
Fig. 3. Examples of typical SEM images of failure modes (100 ) (A) Adhesive failure; (B) Cohesive failure in resin composite; (C) Cohesive failure in resin cement; (D) Mixed failure (C–C model); (E) Mixed failure (C–R model).
Fig. 4. Ceramic surface with or without HF etching (5000 ) (A) Ceramic surface polishing without HF etching; (B) Polished ceramic surface followed by HF etching.
Fig. 5. Elements identified by EDX for Clearfil SA Cement (A) and Variolink II (B).
model. Variolink II showed significantly higher bonding potential to glass ceramics than Clearfil SA Cement (Table 2 and Fig. 1). Although the elements from the two resin cements are similar by
showing high concentration in silicon and barium (Fig. 5), the mean diameter of the filler particles in Variolink II is smaller than that in Clearfil SA Cement, 0.7 μm and 2.5 μm respectively. Resin
T. Tian et al. / International Journal of Adhesion & Adhesives 54 (2014) 165–171
cement with smaller filler particles may be adapting more intimately to the roughened surface and thus result in improved bonding. Another possible reason is the silane that was used in this study. Variolink II exhibited better bonding performance when used with the silane (Monobond S), possibly as these two materials are produced from the same manufacturer and thus highly compatible. A recent study also showed that Variolink II performed better than Clearfil SA Cement when using Monobond S [2]. However, it is unclear if Variolink II or Clearfil SA Cement would achieve the same performance if a different silane coupling agent was applied. Further study is needed. The conventional surface treatment on ceramics showed better bonding behavior in both bonding models than the one recommended in the previous study mentioned (Table 2). This is quite apparent in the groups using Variolink II resin cement. The possible reason might be related to the different ceramic structures used in this study compared with the previous work. In the previous study, a modified leucite reinforced ceramic was with silane application without HF etching. In the present study, however, a lithium disilicate ceramic was used with HF etching followed by silanization. This may indicate that micromechanical attachment created by HF etching is still important in ceramic surface preparation to achieve good adhesion. The first hypothesis that there is no difference in microtensile bond strength between two bonding substrates was rejected. The second hypothesis that there is no difference in microtensile bond strength among different surface treatments and resin luting cements was also rejected.
5. Conclusions Comparing the two bonding models, it is clear that ceramic substrate yields some advantages over the resin composite substrate by providing more reliable and consistent results. However, considering its complex situation in mixed failure, the clinical relevance should be more carefully interpreted. This model is probably more useful for evaluating adhesion of resin materials to ceramics. Consequently, it is necessary to further the improvements on the methodology of ceramic bonding evaluation.
Acknowledgment The authors are grateful to the following companies for providing the experimental materials used in this study: Ivoclar Vivadent, Kuraray Noritake Dental and Dental Ventures of America.
171
References [1] Albakry M, Guazzato M, Swain MV. Biaxial flexural strength, elastic moduli, and X-ray diffraction characterization of three pressable all-ceramic materials. J Prosthet Dent 2003;89:374–80. [2] Bahr N, Keul C, Edelhoff D, Eichberger M, Roos M, Gernet W, et al. Effect of different adhesives combined with two resin composite cements on shear bond strength to polymeric CAD/CAM materials. Dent Mater J 2013;32:492–501. [3] Boscato N, Della Bona A, Del Bel Cury AA. Influence of ceramic pre-treatments on tensile bond strength and mode of failure of resin bonded to ceramics. Am J Dent 2007;20:103–8. [4] Burrow MF, Thomas D, Swain MV, Tyas MJ. Analysis of tensile bond strengths using Weibull statistics. Biomaterials 2004;25:5031–5. [5] Della Bona A, Anusavice KJ, Mecholsky Jr. JJ. Failure analysis of resin composite bonded to ceramic. Dent Mater: Off Publ Acad Dent Mater 2003;19:693–9. [6] Frankenberger R, Taschner M, Garcia-Godoy F, Petschelt A, Kramer N. Leucitereinforced glass ceramic inlays and onlays after 12 years. J Adhes Dent 2008;10:393–8. [7] Hayashi M, Tsuchitani Y, Kawamura Y, Miura M, Takeshige F, Ebisu S. Eightyear clinical evaluation of fired ceramic inlays. Oper Dent 2000;25:473–81. [8] Hernandez AI, Roongruangphol T, Katsube N, Seghi RR. Residual interface tensile strength of ceramic bonded to dentin after cyclic loading and aging. J Prosthet Dent 2008;99:209–17. [9] Ho G, Matinlinna JP. Evaluation of the microtensile bond strength between resin composite and hydrofluoric acid etched ceramic in different storage media. J Adhes Sci Technol 2011;25:2671–85. [10] Hooshmand T, Rostami G, Behroozibakhsh M, Fatemi M, Keshvad A, van Noort R. Interfacial fracture toughness of different resin cements bonded to a lithium disilicate glass ceramic. J Dent 2012;40:139–45. [11] Hooshmand T, van Noort R, Keshvad A. Bond durability of the resin-bonded and silane treated ceramic surface. Dent Mater: Off Publi Acad Dent Mater 2002;18:179–88. [12] Horn HR. Porcelain laminate veneers bonded to etched enamel. Dent Clin North Am 1983;27:671–84. [13] J Symynck, F.d.B., Monte Carlo pivotal confidence bounds for Weibull analysis, with implementations in R; May 13–14 2011. [14] Matsumura H, Kato H, Atsuta M. Shear bond strength to feldspathic porcelain of two luting cements in combination with three surface treatments. J. Prosthet. Dent. 1997;78:511–7. [15] Michelini FS, Belser UC, Scherrer SS, De Rijk WG. Tensile bond strength of gold and porcelain inlays to extracted teeth using three cements. Int J Prosthodont 1995;8:324–31. [16] Pelogia F, Valandro LF, Brigagao V, Neisser MP, Bottino MA. Resin microtensile bond strength to feldspathic ceramic: hydrofluoric acid etching vs. tribochemical silica coating. Int J Prosthodont 2007;20:532–4. [17] Pisani-Proenca J, Erhardt MC, Valandro LF, Gutierrez-Aceves G, BolanosCarmona MV, Del Castillo-Salmeron R, et al. Influence of ceramic surface conditioning and resin cements on microtensile bond strength to a glass ceramic. J Prosthet Dent 2006;96:412–7. [18] Pollington S, Fabianelli A, van Noort R. Microtensile bond strength of a resin cement to a novel fluorcanasite glass-ceramic following different surface treatments. Dent Mater: Off Publ Acad Dent Mater 2010;26:864–72. [19] Roos M, Stawarczyk B. Evaluation of bond strength of resin cements using different general-purpose statistical software packages for two-parameter Weibull statistics. Dent Mater: Off Publ Acad Dent Mater 2012;28:e76–88. [20] Sakai M, Taira Y, Sawase T. Silane primers rather than heat treatment contribute to adhesive bonding between tri-n-butylborane resin and a machinable leucite-reinforced ceramic. Dent mater J 2011. [21] Soderholm KJM. Critical evaluation of adhesive test methods used in dentistry. J Adhes Sci Technol 2009;23:973–90. [22] Stawarczyk B, Ozcan M, Trottmann A, Hammerle CH, Roos M. Evaluation of flexural strength of hipped and presintered zirconia using different estimation methods of Weibull statistics. J Mech Behavi Biomed Mater 2012;10:227–34. [23] Tyas MJ, Anusavice KJ, Frencken JE, Mount GJ. Minimal intervention dentistry —a review. FDI Commission Project 1–97. Int. Dent. J. 2000;50:1–12.