During the physiological function of chewing, or the non-physiological function of bruxism, compressive, bending, and torsional stresses are generated in teeth or prosthetics. These stress will cause the abutment screw loosening or fracture, fracture of the abutment, fracture of the implant, and the implant/abutment connection. The extent of the damage is influenced by the design of the prosthesis, the fit of the implant prosthesis, the implant inclination, and the loading force. In these circumstances, tolerable enough clinical implant/abutment joint strength is required and ISO 14801 is used as their fatigue strength test method. However, torsional stresses generated in the oral cavity should also be considered. For these reasons, this study was undertaken in order to compare the torsional strength required to deform the implant-abutment connection for various diameter implants. The mode of failure will be the future investigation to be resolved by observing the fractured surfaces. In all specimens tested in this study, the relationship between static torsional load and fracture occurrence followed a parabolic curve. The primary curve was determined to be the torsional torque under which plastic deformation occurred and subsequently proceeded, resulting in permanent deformation. The curves illustrated in Fig. 2 showed two patterns in all specimens. Smaller diameter (3.3 and 3.8 mm) implants fractured much easier and earlier at the implant-abutment interface. This load-displacement curve is similar to the result of statistic loading test that Huang HM et al. reported.

Examining all SEM images of the implant-abutment connections, we found that although their anti-rotational notches had been destroyed, the corresponding internal grooves remained intact. When torsion was applied, the grooves, which are composed of grade 4 commercially pure titanium (CP-Ti), were compressed; however, the abutment interlocks, which are composed of a titanium alloy (Ti-6Al-4V), were completely sheared off. Although the tensile strength of the Ti-6Al-4V alloy is at least 2 % greater than that of CP-Ti, much more titanium is adjacent to the grooves, which are compressed and can therefore withstand the transmitted force. On the other hand, the interlocks contain less material and therefore shear off if the applied force is too high. Nagel et al. studied the implant-abutment connection of each of the Replace-Select and the CAMLOG implants using FEM and reported that each design was very similar. When comparing each system, they reported that the Replace-Select implant may fail by fracture of the implant body at the thinnest part of the wall. This thin portion represents the internal design that prevents the implant-abutment from rotating.