The bond strength, that the orthodontic adhesive system is able to generate, influences the outcome of the bracket bonding to the surface of the dental enamel. This is essential for the success of the treatment since, from a clinical point of view, phenomena of accidental debonding could be the cause of damage to the dental enamel, increase the number of appointments necessary and/or extend both the operating times and those necessary to complete the orthodontic therapy [13]. Specifically, the incidence of the detachment phenomena was described both at the enamel—adhesive system interface and at the adhesive system—bracket base interface. It has been observed that these phenomena depend on the value of the bond strength of the orthodontic adhesive, but are influenced also by high mechanical stresses that occur during the orthodontic therapy, or as a result of a decrease in the bond strength at the interface, as occurs when using brackets in polymeric material [8].
F.L. Romano evaluated in vivo the failure rate of adhesion of metal brackets to the tooth enamel surface of both arches with the TXT adhesive system. It was found to be equal to 1.57%, i.e. only 3 brackets out of 190 in a 6-month period underwent accidental debonding [14]. These data are indicative of why this material is, to date, the most widely used adhesive system, as a control, both in clinical and laboratory studies and why TXT has been considered in our in vitro study, despite the abundant scientific literature already provided on it [15,16,17,18,19,20,21,22].
In order to investigate the orthodontic bond strength, the majority of researchers used the strength of the "shear" bond rather than tension or torsion, as the former was found to be the most reproducible. As regards the values of the orthodontic bond strength, reported in the literature, it is actually observed that they can vary considerably also depending on the specimen preparation specifications and the tests conditions [23,24,25,26,27,28,29,30,31,32,33,34].
C. Sturdevant has shown that, depending on the material of which the bracket is made of, the value of the bond strength of the adhesive system, obtained following mechanical tests of resistance to shear strengths, can vary between 17 and 24 MPa [35]. In this regard, T.R. Katona has conducted extensive researches demonstrating that this simple approach of measuring the strength of adhesion, can provide often incorrect results [36]. Indeed, the stress produced on the bracket and on the enamel is not homogeneous, but it is concentrated, generating a local stress greater than that created between the applied force and the interface. The result is an underestimation of the local stress, that causes the failure of the adhesive bond, caused by micro-cracks propagation through the adhesive itself, which is more fragile and, probably, also through one or both interfaces [37, 38].
N. Fox hypothesized an existing relationship between the force application site and the surface of the base of the bracket, noting that a variability in the arrangement of this site and the relative position of the constituents of the adhesion system (enamel-adhesive-bracket), is able to determine substantial differences in the measurement of that force which is responsible for the failure of the adhesive bond [39].
Results of tests with tensile or shear strength can determine a coefficient of variation or relative standard deviation [(standard deviation/mean) × 100%] ranging from 20 to 30%. Typically, tensile force tests produce a lower coefficient of variation than the more common shear strength analysis [8].
By comparing the data obtained in our study, it is possible to highlight how all the adhesive systems under examination provide average SBS values in the range of 3—4.5 MPa. As shown in Table 4, TXT, with an average SBS value equal to 4.45 MPa, was found to be the adhesive material with the greatest adhesion strength, on the contrary Leone proved to have the lowest average SBS values, in fact equal to 3.14 MPa. The Bisco, on the other hand, showed intermediate SBS average values, equal to 4.08 MPa, but with larger variations from the lowest (1.34 MPa) to the highest (7.38 MPa) measurement among all the samples examined (Table 4).
However, it is important to underline that all the data collected by us are far from the parameters suggested by I.R. Reynolds as clinically acceptable levels of adhesion strength, i.e. 6–8 MPa [40]. Various studies have suggested bond strengths between 2.8 and 10 MPa to be clinically adequate [4, 5]. The same comparison, concerning the TXT and the SBS parameters present in the literature, shows an important difference in the results obtained in our study (Table 4) [41]. This significant discrepancy in the results is most likely explained by the fact that our study was purposely conducted without the use of brackets since, from the literature it is clear that the differing geometries of the bases of the bracket are able to greatly influence the strength of adhesion of orthodontic adhesive systems [41]. However, the purpose of our study was to compare the adhesion capacity of the three orthodontic adhesive systems in question and for this reason we chose to isolate the "bracket base" variable, focusing attention on the real resistance that the resin alone offers shear forces.
The objective of many studies in the literature is to demonstrate, following the detachment of the bracket, at what level the breaking of the bond of the bracket to the tooth surface occurs. When testing for adhesive bond failure, three situations may arise: bond failure at the base bracket/adhesive interface, at the adhesive/enamel interface, or cohesive failure [42]. The literature shows that the fracture gap can be localized, to the same extent, both at the enamel/adhesive interface, and at the bracket/adhesive interface [43,44,45,46,47,48,49].
The orthodontic adhesive resin however remaining on the enamel surface, after the bracket debonding, is necessarily mechanically removed from the tooth surface by milling and this entails the risk of accidentally removing even the most superficial layer of the dental tissue [25].
According to S. Elekdag-Turk, the prevalence of adhesive bond failure at the bracket/resin interface becomes a protective phenomenon for the enamel, precisely because, at the moment of detachment, this structure remains intact, preventing damage such as the loss of superficial tissue fragments therefore, cleaning the dental surface from the residual adhesive resin is perhaps less risky than the damage induced by bracket debonding [50]. Regarding the breaking of the bond at the enamel/adhesive interface, according to T. De Melo, the anatomy, the curvature and the design of the base of the bracket, may be the factors responsible for the greater strength of the bonding at bracket/adhesive interface and is just this condition to favour the major preservation of the dental enamel, because only a thin layer of residual material remains to be removed by milling [51].
It is also true that the removal of a bracket attached to the tooth through a high SBS adhesive system can increase the incidence of fractures or micro-cracks affecting the enamel, because of the increased effort have to be spend to remove the retained adhesive resin from the enamel surface [25]. Actually, what is considered as desirable in the debonding milling manoeuvres is to remove the bracket without damaging the enamel surface. It was also quantified that even the safest debonding manoeuvres were found to be responsible for the loss of about 10–20 µm of surface enamel [8].
The greater SBS to dental enamel demonstrated by TXT, could increase the probability of iatrogenic lesions of the hard tissues of the tooth occurring during the manoeuvre of bracket debonding. On the other hand, the greater affinity for chemical bonding, or cohesion, of TXT for the metal surface of the base of the bracket, would guarantee the orthodontist better holding performance of the bracket, during the active phases of the treatment.
The FE-SEM analysis (Fig. 2) of the bracket bases coupled at one of the three orthodontic adhesive systems under examination, has identified the adhesive interface in which the fracture and detachment of the bracket most commonly occurred. As regards TXT, the fracture occurs at the enamel-adhesive interface, given the large amount of resin residual on the base of the bracket (Fig. 2a). Highly filled resin composites have been observed to bond to mechanically retained metal brackets better than lightly filled composites [52]. Which once again TXT is the material repeatedly subjected to tests, some of the studies in the literature presented showed a tendency for the adhesive bond to break mainly at level of the enamel-adhesive interface, rather than the adhesive-bracket one [25, 27, 28, 30, 53]. This latter evidence is precisely in accordance with the results obtained in our in vitro study.
It has been reported that, enhancement of bond strength may compromise safe bonding, in fact, the detachment of a large part of TXT from the dental surface could expose, even if in a minimal percentage (1–30%), the enamel to damages such as micro-fractures or loss of superficial hard tissue. Bisco and Leone, which have instead shown a greater tendency to remain adherent to the enamel surface, have to be mechanically removed from the tooth by mechanical milling, taking care not to damage the dental enamel. It has been observed that, the resin tags that remain on the enamel surface, for a long time after the bracket debonding, can change colour as well as constitute sites for bacterial adherence [54]. Finally, the FE-SEM images however show a good performance of the Ovation bracket in most of the samples examined.
The investigation of aging in saliva and in sugary drink has been performed with the intent of establishing whether the recorded dimensional variations, even if minimal, were still capable of altering the physical–chemical properties of the surface of the three orthodontic composite resins under examination. Data on weight changes in saliva, Fig. 3a, showed a small weight increase in the order of 10−4 g, which is however considered not significant in terms of the chemical stability of the examined materials. These minimal modifications in weight values can be expected and caused by the surface adsorption phenomenon of organic saliva residues, proteins and mucus, as the result of reversible interactions, as it does not imply any type of irreversible chemical reaction as can be seen also by the unchanged colour of the sample surface. Overall, this enhancement in weighting resulted not uniform in the trends belonged to each orthodontic adhesive resins, although it showed a tendency to stabilize over time. The investigation of ageing in saliva, it has been followed that the minimal dimensional variation recorded in the tested orthodontic resins does not able to alter physical–chemical properties in the surface.
As showed in Fig. 3b, which summarizes the weight measurements of the samples over 30 days, even in this case the trends of Bisco and Leone resins can be overlapping, a similar trend for them is shown, while TXT reveals a weight saturation after fifteen days, thus demonstrating a different type on interaction with the testing solution. About weight changes for samples stored in sugary drink, it is showed a similar behaviour than which one recorded in saliva: a different interaction of the TXT is recorded both in saliva than in the sugary drink respect Bisco and Leone adhesive resins.
Overall, the weight changes of the three materials are again considered almost negligible and no materials losses can be evaluated in all the tests; therefore, all the orthodontic resins have shown to be resistant to the acidic and corrosive components of the sugary drink used in the experiment.
Raman Spectroscopy was a powerful tool, it has been used to investigate on the chemical composition of the three orthodontic adhesive resins, especially on the nature of their dispersed phase and, to establish the effects that occur following phenomena of aging in saliva and sugary drink, precisely at the level of the surface. Therefore, the purpose of this analysis has been twofold: a simple compositional comparison of the three adhesive resins (Fig. 3b) and an evaluation of the quality of the three materials (Fig. 4) in response to an exposure, until one month, in two different aging solutions to likely correlate these findings to possible changes in mechanical behaviour of the materials. On the basis of the chemical bonds reproduced in the Raman spectra it has been established that the type of filler prevalent in the dispersed phase of each resin is quartz. Strong affinities between the spectra of Bisco and Leone have been revealed, while the TXT exhibited a peculiar peak at 450 cm−1 associated to the higher percentage of quartz in its chemical composition (Fig. 4), as also reported by its safety data sheet. This last data allowed to justify both the diversity in the adhesive behaviour of the TXT resin and the similarities of performances demonstrated by the Bisco and Leone resins and finally, to justify the greater chemical affinity, or cohesion force which, following debonding, the TXT showed (Fig. 2a) with the metal interface represented by the base of the Ovation bracket, compared to the other two orthodontic resins tested (Fig. 2b, c). It is known that a higher content in fused quartz fillers, involves the attainment of high compressive strength and stiffness, the abrasion resistance and, the reduction the thermal dimensional change of the resin to a value matching that of tooth structure, effectively increasing adhesion to both the interfaces: adhesive/enamel and adhesive/base bracket. The surface profile and microstructure of the orthodontic composites are subjected to changes arising from degradation and wear processes in service; through Raman analysis it was observed that, the ageing in saliva usually produces the appearance of new broad bands in the region ranging from 900 to 2500 cm−1 ascribed to a possible fluorescence signal coming from the adsorption of organic residues [55]. Here, all the three materials demonstrate exhibit a good behaviour showing small changes in the spectra without the presence of pronounced new bands, thus indicating the absence of relevant and permanent surface effects (Fig. 5). Conversely, in case of ageing in sugary drink, only Bisco does not reveal any variation in its spectrum. Therefore, Raman analysis on samples stored in saliva and sugary drink highlighted that, Bisco is weakly contaminated with respect to the other two materials, and we speculate that this could be associated to peculiar moisture resistance properties.