P. Pallav, PhD

This phenomenon occurs in moving contacts and/or at cyclic load deviations.
Due to normal (dental) contact geometry, contact stresses concentrate slightly below the surface. (explained here)
In the case of surface fatigue a subsurface region is (or has been) exposed to very high multi-directional stresses which alternate within a few milliseconds if the contact moves. Because of this, a network of subsurface cracks is created. If, ultimately enough cracks extend to the surface, fragments break out and leave a very rough, pitted surface.

Surface fatigue is generally inversely proportional to the hardness (BHN, VHN), which follows from contact mechanics. Irregularly shaped inclusions (filler particles in composites) induce stress concentration and enhance fatigue phenomena.
In tribology, surface fatigue is not always considered a wear mechanism. The more dramatic the surface fatigue, the closer it gets to a mechanical overload, rather than wear.

The deformations are exaggerated to illustrate what happens. When the antagonist slides as drawn, each square at the highly loaded second row in the figure will go through all stages of deformations in that row. This works like a reverse bending fatigue test.

Example

A gold alloy cusp (E ≈ 75 GPa) with a (spherical) radius of curvature of 5 mm at the 1 mm long area of contact, pressed against a plane surface of composite (E ≈ 20 GPa) with a (static) force of just 1 N, results in a contact pressure of about 95 MPa(!) on a 40 µm wide contact line. If, in addition a certain amount of impact is present, the contact force may temporarily be much higher than the static value and the associated contact stresses should be calculated accordingly.

The quotient,

S2fatigue E	where Sfatigue =	Fatigue Strength
	E =	Elastic module

can be used to pre-select materials on their sensitivity to surface fatigue in initial line contacts. This is Equation 2 in Contact Stress solved for the normal force. Note the resemblance of this expression with the resiliency and toughness.

When a random set of materials is tested for surface fatigue and for the elastic modules, the outcome will very likely be that materials, which have high elastic module are more resistant to surface fatigue. Although this appears to be in contrast with the appearance of E in the denominator, this is simply because with many materials there is a strong linear correlation (proportionality) between S_fatigue and E (and other properties).

In view of the forces calculated for the contact between the gold alloy cusp and the composite plane, overload at direct contacts is probably inevitable.