Methods: Twelve intact freshly extracted human third molars were used for µTBS. Optibond FL (OBFL, Kerr) served as control. After bonding procedures following the respective manufacturer's instructions and subsequent composite build-up (Z100, 3M ESPE), specimens were sectioned to 1.8x1.8 mm sticks and trimmed at the interface to a cylindrical hour-glass shape (±1.2 mm) using the MicroSpecimen Former (University of Iowa). TEM non-demineralised and demineralised sections, and sections of specimens immersed in ammoniacal silver nitrate were prepared by ultra-microtomy and examined (EM 900, Zeiss).
Results:
| µTBS | CMF | OBFL | p-value (t-test) |
| Enamel | 30.7 ± 9.2 MPa, n=10 | 26.7 ± 5.0 MPa, n=11 | p=0.3512 |
| Dentin | 25.7 ± 5.8 MPa, n=10 | 45.6 ± 9.9 MPa, n=10 | p<0.0001 |
No pre-testing failures were recorded.
TEM showed a tight interface at enamel. The etch-pits produced by the 'mild' phosphoric-acid conditioner (pH=1.5) were rather shallow, but resulted in clearly defined micro-tags. A relatively thick, completely demineralised and acid-resistant hybrid layer was formed, varying in thickness between 1.5-3 µm at dentin. The dentin tubules were opened and enabled the formation of resin tags. Some degree of spot and cluster-like nanoleakage was observed, but did not differ in extent and form in comparison to that observed with the OBFL control.
Conclusion: The new three-step adhesive 'cmf adhesive system' showed clear indications of good bond reliability.