During polymerization of resin composites, shrinkage stresses compete with
resin-dentin bonds in a manner that can cause failure of the bond, dependin
g upon the configuration of the cavity, its depth, and the restorative tech
nique. The hypothesis tested in this study was that the effect of cavity co
nfiguration (C) and remaining dentin thickness (RDT) influence resin bond s
trength to the dentin of Class I cavity floors. The occlusal enamel was gro
und to expose a flat superficial dentin surface as a control (superficial d
entin, C-factor = 1) in human extracted third molars. Cavities 3 mm long x
4 mm wide were prepared to a depth 2 mm below the ground dentin surfaces (d
eep dentin within cavity floor, C-factor = 3). To assess the relationship b
etween C-factor and RDT, we removed the walls of cavities, making a deep fl
at surface for bonding (deep dentin, C-factor = I). The teeth were restored
with either Clearfil Liner Bond II (LB II), One-Step (OS), or Super-Bond D
Liner (DL), followed by Clearfil Photo Posterior resin composite. After 24
hrs' storage in water, the teeth were sectioned vertically into 3 or I sla
bs (0.7 mm thick) and trimmed for the micro-tensile bond test so that we co
uld determine the strength of the resin bonds to the pulpal floor. All grou
ps gave high bond strengths to superficial dentin, but OS and DL gave signi
ficantly lower bond strengths to flat deep dentin when the C-factor was 1.
When the C-factor was increased to 3 by the creation of a three-dimensional
cavity preparation, the bond strengths of ail materials fell (range, 21 to
35%), but the difference was significant (p < 0.05) only with DL. SEM obse
rvations of failure patterns showed that specimens with high bond strengths
tended to exhibit cohesive failures within the hybrid layer, while specime
ns exhibiting low bond strengths showed failures at the tap of the hybrid l
ayer. Some adhesives do not bond well to deep dentin, making them more susc
eptible to polymerization shrinkage stress that develops in cavities with h
igh C-factors.