Theory of “Cohesive” vs “Adhesive” Separation in an Adhering System

Abstract It is shown that there is limited validitity to the doctrine that true interfacial separation, in an adhering system, is highly improbable. An analysis employing the Griffith-Irwin crack theory yields these results: The important parameters are, difference in elastic moduli, ΔE; differences in g, the energy dissipation per unit crack extension; thickness, Δ1 or δ2, of the region where dissipation occurs; and the presence or absence of strong interfacial bonds. If the forces across the interface are appreciably weaker than the cohesive forces in either phase, there is a strong minimum in g at the interface. For flaws of equal size, an interfacial flaw will be the site of initiation of failure. If strong interfacial bonds are present, then if Δg and ΔE have the same sign, failure is most probable, deep within one phase. If Δg and ΔE have opposite signs, failure may be initiated, and may propagate, at a distance δ from the interface, in the phase with lower g. This may be mistaken for weak-boundary layer failure.