Why interface testing by single-fibre methods can be misleading

Abstract

The reasons for our interest in the interface are reviewed, and bond-strength sensitive properties are discussed. It is shown that the four direct methods currently used for interface measurement have problems that make their results inconsistent between methods and inaccurate. The failure process has been monitored in glass/epoxy pull-out specimens with very short fibre free lengths, so that crack propagation is stable. This showed that, close to ultimate failure, a very small length of the interface was carrying nearly all the (very high) load. This apparently high interfacial shear strength can be explained on the basis of the recently documented lack of a shear failure mode for most polymers. All four interface tests are centro-symmetric, or nearly so. In this situation, a matrix that cannot fail in shear, must fail in tension. Yet tensile failure is strongly inhibited by the centro-symmetry. Hence shear failure is forced at the interface itself by a tensile-type process. This involves tensile pulling of the polymer over the high points on the fibre. (No surface is perfectly smooth.) The result is thus unduly influenced by the surface roughness of the fibre and the shrinkage pressure, neither of which are directly related to adhesion. Although great progress has been made with fracture analysis of interface failure, this cannot resolve the problem of the unrealistic centro-symmetry of the tests. It is, therefore, recommended instead that direct tests on composites be examined for relevance and usefulness. The most useful method may well be to measure the off-axis tensile strength of unidirectional composites. This has been shown to be more sensitive to interface properties than both the fragmentation test and the pull-out test. Single-fibre tests and fracture analyses of these tests should be used only for basic research into centro-symmetric failure modes.