Microtensile Bond Strength and Interfacial Characterization of 11 Contemporary Adhesives Bonded to Bur-cut Dentin

Current adhesives can be categorized by adhesive approach and clinical application procedure.1 According to this classification, adhesives belong to either the etch&rinse group, if a separate etchant is applied and rinsed off, or to the self-etch group, if an acidic monomer is used to demineralize and infiltrate the tooth surface simultaneously. Further categorization is based on the number of clinical application steps. The more conventional adhesives have a separate hydrophobic bonding resin that involves an additional clinical application step and are referred to as three-step etch&rinse and two-step self-etch adhesives. The alternative adhesives combine the priming and bonding functions to reduce the number of clinical steps, resulting in the two-step etch&rinse and one-step self-etch adhesives. All these simplifications have repercussions on bonding durability,2 at least for the older generation of simplified adhesives. Clinically, the most attractive are the self-etch adhesives, because an additional rinse and drying step is no longer needed. This reduces the risk of contamination on the surface, eliminates over- and under-drying issues3 and does not require the cotton rolls to be refreshed (in case no rubber dam is used). With respect to bond strategy, the concept of simultaneously demineralizing and infiltrating the tooth surface is advantageous,4 though so far, hydrophilic resins have been employed that make the bond sensitive to hydrolytic degradation.5 

As adhesive technology is rapidly evolving, on the dental market, commercial adhesive formulations are replaced frequently. As a result, many adhesives that are available today have no independent clinical validation for their use. Therefore, laboratory-screening tests remain indispensable in providing data that, to a certain degree, predict clinical effectiveness. Although a direct correlation between laboratory and clinical research has not yet been shown, a clear trend exists that adhesives that present repeatedly and reproducibly with relatively high bond strengths and appear resistant to diverse forms of “aging” also present with high retention rates in clinical Class-V studies.2,6–9 Currently, the best screening method is a combination of quantitative bonding effectiveness measurements and knowledge of the interfacial interaction of new adhesives. Therefore, mechanically and ultra-morphologically, the current study evaluated different contemporary adhesives bonded to dentin. A three-step etch&rinse (OptiBond FL) and a two-step self-etch (Clearfil SE Bond) adhesive, both of which have repeatedly been shown to be excellent performers in clinical and laboratory studies, served as “gold-standard” adhesives for their class. The hypothesis tested was that recently launched, simple-to-use adhesives have a similar bonding effectiveness to dentin as control gold-standard multi-step adhesives.

Eleven adhesives, all currently available on the dental market, including two three-step etch&rinse, three two-step etch&rinse, two two-step self-etch and four one-step self-etch adhesives, were chosen (Table 1). Fifty-five sound human molars (33 for μTBS and 22 for TEM), gathered following informed consent approved by the Commission for Medical Ethics of the Catholic University of Leuven, were used. The teeth, stored in a 0.5% chloramine solution, were used within three months of extraction. Flat dentin surfaces were prepared by removing the coronal tooth part with an Isomet low-speed diamond saw (Isomet 1000, Buehler, Lake Bluff, IL, USA). Then, a standardized smear layer was prepared using a regular-grit diamond bur (842, Komet, Lemgo, Germany) mounted in the MicroSpecimen Former (University of Iowa, Iowa City, IA, USA). The dentin surfaces were verified for the absence of enamel and/or pulp tissue exposition using a stereo-microscope (Wild M5A, Heerbrugg, Switzerland).

Three teeth were used per adhesive. The adhesives were applied strictly following the manufacturer's guidelines (Table 1). A composite build-up was then made using a single resin composite (Z100, 3M ESPE, St Paul, MN, USA), which was applied in five increments with a height of approximately 1-mm and light-cured for 40 seconds with an Optilux 500 light-curing device (Demetron, Kerr, Danbury, CT, USA) with a light output of not less than 550 mW/cm².

After 24-hour storage in water, the resin-dentin bonded specimens were sectioned with a water-cooled diamond saw (Isomet 1000, Buehler Ltd, Lake Bluff, IL, USA) in both the x and y directions to obtain rectangular sticks (six to seven) from the central part of the coronal dentin surface. The dimensions of the sticks were then measured by means of a digital caliper (CD-15CPX, Mitutoyo, Kanagawa, Japan) from which the cross-sectional area was calculated (approximately 0.9 mm²). The non-trimmed micro-specimens8 were fixed to a modified microtensile bond testing jig8 with cyanocrylate glue (Model Repair II Blue, Dentsply-Sankin, Tokyo, Japan) and tested in tension at a crosshead speed of 1.0 mm/minute using an LRX testing machine (Lloyd, Hampshire, UK) equipped with a load cell of 100 N. The bond strength values were calculated in MPa by dividing the imposed force (in N) at the time of fracture by the bond area (in mm²).

The failure modes were evaluated with a stereomicroscope (Wild M5A) at a magnification of up to 50× and categorized as “interfacial,” “cohesive” (dentin or composite) or “mixed.” The data were statistically evaluated by one-way ANOVA and Tukey's Multiple comparisons test at a significance level of α=0.05.

The bonding mechanism to dentin was morphologically assessed by TEM (JEM-1200EX II, JEOL, Tokyo, Japan). Two dentin surfaces were prepared for each adhesive in the same way as for μTBS testing. Following adhesive treatment, the resin-bonded dentin specimens were cross-sectioned perpendicular to the resin-dentin interface to obtain 0.8-mm wide sticks using the slow speed diamond saw. Half of the specimens were then demineralized and fixed simultaneously in a 10% formaldehyde-formic acid solution (Gooding and Stewart Fluid, Prosan, Gent, Belgium) for at least 36 hours. TEM sample preparation of both the demineralized and non-demineralized sections was then performed in accordance with common procedures used for ultra-structural TEM examination of biological tissues. 11 Eventually, 70-90-nm thick sections through the resin-dentin interface were cut using a diamond knife (Diatome, Bienne, Switzerland) in an ultramicrotome (Ultracut UCT, Leica, Vienna, Austria).

The μTBS of the 11 adhesives to dentin are shown in Table 2 and Figure 1. The mean μTBS values ranged from 11 to 63 MPa. The lowest value was obtained for the one-step self-etch adhesive (Adper Prompt L-Pop) and the highest was obtained for the three-step etch&rinse adhesive (OptiBond FL). The one-step self-etch adhesive tended to have lower μTBS values than the two-step self-etch and etch&rinse adhesives, though a product-related dependency also existed. The three-step etch&rinse adhesive OptiBond FL had not only the highest μTBS, it was also statistically different from all the others. Pre-testing failures (mainly failures during sectioning with the slow-speed diamond saw) occurred only with the one-step self-etch adhesive Adper Prompt L-Pop (4 out of 18). No pre-testing failures were recorded for any other adhesive tested.

Failure analysis showed interfacial, cohesive and/or mixed fractures, depending on the adhesive tested (Table 2). A general trend was observed: specimens that presented with lower bond strength failed more at the resin-dentin interface (interfacial failure). On the other hand, specimens with higher bond strengths failed more cohesively in dentin or resin.

Regarding self-etch adhesives, the interfacial-ultra-structure differed, depending on the actual self-etch approach and, thus, the specific adhesive, especially when the degree of interaction at the interface was more intense with adhesives that provided a relatively low pH primer. The one-step self-etch adhesive Adper Prompt L-Pop is a strong self-etch adhesive, since its interaction with dentin was very similar to that of etch&rinse adhesives. A thick (3–5 μm) hybrid layer, wide resin tags and removal of all residual hydroxyapatite within the hybrid layer was observed (Figure 2a). The one-step self-etch adhesive Xeno III interacted somewhat less intensely with dentin; the hybrid layer was about 2 μm thick and, in the bottom part, some hydroxyapatite was preserved (Figure 2c). The two-step self-etch adhesives Clearfil SE (gold standard) and Protect Bond are typical mild (pH of the primer is around 2) self-etch adhesives; the hybrid layer was less than 1 μm thick and only partially demineralized over its full width (Figures 2e and 2f). The remaining one-step self-etch adhesives G-Bond and Clearfil S³ Bond were typical ultra-mild self-etch adhesives; their interaction with the underlying dentin was rather superficial (100-200 nm) and cannot be distinguished from a potentially resin-impregnated smear layer (Figures 2c and 2d).

Regarding etch&rinse adhesives, the interfacial-ultra-structure was very similar for all adhesives (Figure 3). As all the adhesives made use of phosphoric acid to initially etch dentin, they presented with thick (3–5 μm), completely demineralized, hydroxyapatite-free hybrid layers with wide open tubules filled with resin tags that extend even into the small lateral tubule branches. Marked differences were observed in the appearance of the adhesive resin itself and, in particular, its filler. Some adhesives had no filler at all, such as Scotchbond Multi-Purpose (Figure 2d), while others contained nanofillers (Prime&Bond NT, Scotchbond 1 XT and XP Bond; Figures 3a, 3b and 3d, respectively) in different densities. The adhesive resin of OptiBond FL (gold standard) was heavily loaded with conventional glass filler, with particles up to several micrometers (Figure 3e).

In the current study, the bonding effectiveness of 11 adhesives to dentin was comparatively investigated. The hypothesis, that recently launched simple-to-use adhesives had a similar bonding effectiveness to dentin as control gold-standard multi-step adhesives, was rejected, as most simplified adhesives, and especially the one-step self-etch adhesives, revealed a significantly lower bond strength than that of the gold-standard three-step etch&rinse control adhesive (Table 2).

In the current study, the authors used the μTBS test to mechanically assess the strength of the resin-dentin interface complex. Today, the μTBS test is one of the most commonly used methodologies, since it has several advantages over the more traditional macro-tensile and shear-bond test methodologies.4,12–13 The μTBS test is more versatile, as, for example, multiple specimens can be obtained from a single tooth, enabling more sophisticated study setups and better controlled substrate variables. Also, the stress is more uniformly distributed during loading across the interface, again as compared to the more traditional bond strength test methods.3,14 Many modifications to this μTBS method have been proposed in the literature,15 especially as the final preparation of the interface area appears important, since cracks can be introduced at this level. Aggressive and insufficiently controlled trimming of these micro-specimens may introduce interfacial defects that may lead to premature failures and a reduction in the observed bond strengths by the early onset of crack propagation during tensile loading of the specimen. Originally, the rectangular constriction at the interface of these micro-specimens was prepared by hand using a dental handpiece. 16 This method is not only laborious, it also largely depends on the skills of the operator, thereby introducing a non-negligible factor of technique sensitivity. Theoretically, circular constriction leads to a better stress distribution at the interface area.17,19 This additional microspecimen milling is best performed using a more-or-less automated device, such as the MicroSpecimen Former (University of Iowa), so that the specimen preparation is better standardized and less operator specific. 20 Also, other variables, such as grip, loading speed and alignment, are very important and should be standardized.4,21 In the current study, the authors opted for non-trimmed μTBS specimens that combine a good stress distribution at the interface17 with a minimal amount of processing; the 1x1 mm resin-dentin sticks are cut out of the restored tooth and directly transferred to the universal testing machine.

Only the central dentin portion that is located directly above the pulp was used in the current study in order to minimize any regional variation between the periphery and central dentin substrate.4,22–23 (This, however, reduced the number of specimens available for testing [6–7, instead of up to 30 in other studies15]) but was thought to increase the validity of the results.

In the currrent study, as in several other studies,2,14 the highest μTBS was obtained with the conventional three-step etch&rinse adhesive OptiBond FL. There was a marked, statistically significant difference between OptiBond FL and the other three-step etch&rinse adhesives tested, namely Scotchbond MP. In addition to a difference in the actual bonding effectiveness, the μTBS also reflects the whole strength of the interface complex. The higher μTBS of OptiBond FL should probably also be attributed to the high filler loading of the adhesive resin and high mechanical strength, as proven by its relatively high e-modulus, comparable to that of several flowable composites.24 On the other hand, the adhesive resin of Scotchbond MP does not contain any filler and has previously been shown to be less effective. 11 Nevertheless, the μTBS of Scotchbond MP remains comparable to that of the two-step etch&rinse adhesives Scotchbond 1 XT and Prime&Bond NT. XP Bond, which is relatively densely filled with silica (even accumulating in the resin tag, Figure 3d) and makes use of ter-butanol as a solvent, scored bond strengths intermediate to the non-filled and highly filled resin. Ter-butanol has a similar vapor pressure to ethanol but without the risk on esterification of functional monomers. XP Bond also performed well when tested following a conventional shear bond strength method test performed by general practitioners in a so-called battle of the adhesives series by Degrange and others.25 

Another issue that is relevant to all etch&rinse adhesives is the manner in which dentin is treated after the etchand-rinse procedure. This is a very technique-sensitive step, where the operator should take care not to “under-“ or “over-dry” the dentin surface, as both are known to result in impaired bonding effectiveness.3,26 In the first place, the window of opportunity to achieve optimal hybridization is dependent on the solvent used in the primer and is considerably smaller for (water-free) acetone-based primers27 than for water/ethanol-based primers—the latter being obviously more forgiving with regard to the degree of dentin wetness.28 The only water-free acetone-based adhesive tested in the current study was Prime&Bond NT, which scored the lowest bond strengths of all the etch&rinse adhesives.

Two-step etch&rinse adhesives combine the primer and adhesive resin in order to reduce the number of clinical application steps from three to two. In the current study, three two-step etch&rinse adhesives (XP Bond, Scotchbond 1 XT and Prime&Bond NT) were tested. Their μTBS were significantly lower than that of the three-step etch&rinse adhesive OptiBond FL, as corroborated by former studies conducted using a similar protocol in the laboratory of the current authors.20,29–30 The immediate bond strength is expected to primarily originate from the micro-mechanical interlocking of resin in the collagen fibril mesh. Therefore, in general, there is no difference in bonding effectiveness of two- and three-step etch&rinse adhesives when tested to flat surfaces and in the short-term. As soon as one tests these adhesives in complex cavities19 or in the longer-term,31 three-step etch&rinse adhesives easily outperform their two-step equivalents.2 

Self-etch adhesives are clinically most attractive, because they do not need a rinse step, thus avoiding the technique-sensitive drying of etched dentin. Self-etch adhesives differ, not only because of the number of clinical steps (two- and one-step self-etch adhesives), but also because of the wide interaction intensity that can be observed. A clear correlation between the pH of self-etch primer and the depth of interaction with dentin was observed in the current study and is graphically summarized in Figure 4. Today, four categories can be distinguished:1,29 1) strong self-etch adhesives have a pH lower than 1 and an interfacial micro-morphology (3–4 μm deep fully demineralized hybrid layers) that is very similar to that of etch&rinse adhesives; 2) they are intermediately strong self-etch adhesives with a pH of around 1.5. These adhesives have a hybrid layer of about 1-2 μm, wherein, at the bottom part, some hydroxyapatite is preserved; 3) they are mild self-etch adhesives with a pH of around 2. The hybrid layer is less than 1 μm thick and is only partially demineralized; 4) recently, a category of ultra-mild self-etch adhesives was added for self-etch adhesives that come with a primer pH higher than 2.5.32 These adhesives do not remove the smear layer and interact with smear layer covered dentin only up to a few hundredths of a nanometer. As interaction with intact dentin is almost non-existent, the micro-mechanical resistance of the interface complex is very dependent on impregnation and stabilization of the smear layer. As a result, the μTBS is far more dependent on the preparation of the surface.33 Therefore, these adhesives bond well to some dentin surfaces but less effectively to others, by which another kind of technique sensitivity is introduced. Clinically, an adhesive that can bond to any surface is preferred, as can mild self-etch adhesives.34 This clear correlation between the pH of self-etch primer and interaction intensity is graphically represented in Figure 4, showing all the key morphological features of the different self-etch approaches.

In light of bond durability, mild and ultra-mild self-etch adhesives have some unique properties that other, more aggressive approaches, such as etch&rinse adhesives, lack. Since not all hydroxyapatite is removed from the interaction zone, much calcium is available for additional chemical interaction with specific adhesive functional monomers.35–36 Some of these bonds are stable, even in an aqueous environment,36 so that the interface can better withstand the hydrolytic breakdown of its components. This mechanism is supposed to prolong the clinical lifetime of restorations.7 

Adper Prompt L-Pop is a strong one-step self-etch adhesive. This adhesive scored the lowest bonding effectiveness of all the adhesives tested (11.1 MPa), and it was also the only adhesive for which pre-testing failures were recorded (4 out of 18 specimens). A more detailed long-term study revealed that, most probably, less optimal polymerization and mono-polymer stability are the basis for these less favorable results.2 This poor in vitro performance is corroborated with several clinical Class-V studies that report less favorable in vivo performance for this strong self-etch adhesive.6,37–38 

Xeno III is a typical intermediately strong self-etch adhesive. The smear layer was completely removed and a 2-μm thick hybrid layer was produced (Figure 2b). Although its μTBS was the third lowest of all the adhesives tested, it was not significantly different from the μTBS measured for several multi-step adhesives, such as Scotchbond MP (Table 2); clinically, up to 10% of the Class V restorations de-bonded after only two years of clinical service, as reported in two randomized Class V studies.18,39 This is considerably worse than the 100% retention rate at five years for the control adhesive OptiBond FL and the 94% retention rate at seven years.8 

Clearfil SE and Protect Bond are very similar adhesives in composition. The most important compositional differences are inclusion of the antibacterial monomer MDPB and sodium fluoride, which are deemed beneficial for the long-term performance of adhesive bonds. The anti-bacterial properties of Clearfil Protect Bond have been proven in vitro,40 but the potential clinical benefit is difficult to assess and, therefore, it remains uncertain. The clinical performance of both adhesives in randomized Class V studies is excellent,9,41 and even approaches that of the gold-standard three-step etch&rinse adhesive OptiBond FL.8 These excellent in vitro and in vivo results must be ascribed to their unique two-fold bonding mechanism related to the mild self-etch approach, which comprises a micro-mechanical bonding component through the formation of a 1 μm thick hybrid layer that may provide resistance to acute debonding stress (as imposed during μTBS testing). Additionally, these adhesives contain the functional monomer 10-MDP, which enables an intensive and stable chemical bond with hydroxyapatite,36 which was left around the collagen fibrils within the hybrid layer (Figures 2e and 2f). Such primary chemical interaction improves the resistance to hydrolytic breakdown42 and, thus, clinically keeps the restoration margins sealed for a longer period.

Also, both of the ultra mild self-etch adhesives that were tested, G-Bond and Clearfil S³ Bond, contain functional monomers with additional chemical bonding potential, suggesting a similar two-fold bonding mechanism. TEM revealed that the interaction with dentin is much shallower (Figures 2c and 2d) and is limited to a few hundred nanometers.32 Therefore, as mentioned above, this twofold bonding mechanism might be compromised for these ultra-mild self-etch adhesives by interference of the surface smear. Short-term clinical evaluation of these adhesives in Class V restorations suggests, however, adequate bonding performance to this lightly prepared, highly mineralized dentin tissue.1,43 Long-term clinical follow-up and adhesion to different substrates, such as carious dentin, are therefore of concern to this adhesive approach and should be investigated in future research.

The bonding effectiveness of current commercial adhesives is not equal. The highest μTBS obtained in the current study was with the three-step etch&rinse adhesive OptiBond FL, which remains the gold standard. Mild two-step self-etch adhesives that also provide additional chemical bonding appear to be the best compromise, as they combine optimal bonding effectiveness with a simplified application protocol.

The authors thank the manufacturers for supplying the materials for this study. KL Van Landuyt has been granted a postdoctoral fellowship by the Research Foundation-Flanders (FWO). This study was supported in part by the FWO No G.0206.07 and KULeuven OT/06/55 research grants.