Why do shear bond tests pull out dentin?

It is widely accepted that a dentin shear bond test which pulls out dentin must mean that the adhesive strength is superior to the cohesive strength of the dentin. Using numerical modeling techniques, Van Noort et al. (1988, 1989) and DeHoff et al. (1995) alerted the scientific community that there were massive stress concentrations in the familiar dentin bond test. It is not inconceivable that these localized high tensile stresses could initiate cracks which diverge monolithically into dentin, leaving the interface unchallenged. To test this hypothesis, we developed a failure accumulation simulation program which determined localized failure interactively "on the fly" with a finite element solver, and also included brittle behavior, adhesive and cohesive failure, stochastic response, and dynamic remeshing. All of the familiar dentin bond variables were included in the simulation. A parallel experimental dentin bond test validation was run, and the fractography was examined in the scanning electron microscope for mode of failure. The simulation confirmed the tensile monolithic fracture hypothesis. It is also confirmed that dentin pull-out was partly due to the biomechanics of the test and did not necessarily mean superior adhesive strength or even that the cohesive strength of the dentin was reduced. There is clear need for a new technology for the evaluation of biological interfaces, and the present work has shown the vital role of numerical modeling in the interpretation of such experimental procedures.