Posterior composite restorative materials undergo accelerated wear in
the occlusal contact area, primarily through a fatigue mechanism. To f
acilitate the timely development of new and improved materials, a pred
ictive wear model is desirable. The objective of this study was to dev
elop a finite element model enabling investigators to predict crack pr
opagation directions in resins used as the matrix material in composit
es, and to verify these predictions by observing cracks formed during
the pin-on-disc wear of a 60:40 BISGMA:TEGDMA resin and an EBPADMA res
in. Laser confocal scanning microscopy was used to measure crack locat
ions. Finite element studies were done by means of ABAQUS software, mo
deling a cylinder sliding on a material with pre-existing surface-brea
king cracks. Variables included modulus, cylinder/material friction co
efficient, crack face friction, and yield behavior. Experimental resul
ts were surprising, since most crack directions were opposite previous
ly published observations. The majority of surface cracks, though init
ially orthogonal to the surface, changed direction to run 20 to 30 deg
rees from the horizontal in the direction of indenter movement. Finite
element modeling established the importance of subsurface sheer stres
ses, since calculations provided evidence that cracks propagate in the
direction of maximum K-II(theta), in the same direction as the motion
of the indenter, and at an angle of approximately 20 degrees. These f
indings provide the foundation for a predictive model of sliding wear
in unfilled glassy resins.