In this paper, we examine the regions of debonding between the fibers and the matrix surrounding fiber breaks formed during single-fiber fragmentation tests. The materials used were E-glass fibers embedded in a matrix of diglycidyl ether of bisphenol-A (DGEBA) cured with meta-phenylenediamine (m-PDA). Tests were performed by imposing sequential increments of strain (step-strains) until no other fiber breaks could be made to occur. At that point, the lengths of the fiber fragments are less than the critical length necessary to transfer sufficient stress to cause additional fiber breaks. That point is termed saturation. The fiber breaks are accompanied by areas of debonding between the matrix and the surface of the fiber. The tensile strain in the matrix adjacent to the debonds can account for a significant amount of the stored strain energy released by the fiber fracture, and this mechanism should be accounted for when analyzing the fragmentation process. With increasing applied strain, the lengths of these debonded regions also generally increase. We contend that the edges of each debond remain fixed to the same location on the fragment surface during loading and unloading. With this assumption, measurement of the lengths of the debonded regions at saturation and again after the specimen have been unloaded suggests that, at saturation, the matrix tensile strain adjacent to the debond regions is an order of magnitude higher than the applied strain (40 % vs. 4 %). Although the edges of the debonds typically remain attached at the same locations on the fiber fragments, debond propagation along fiber fragments under increasing strain has been observed in some cases. This phenomenon is termed secondary debond growth, and two mechanisms that trigger secondary debond growth have been observed. Secondary debond growth can be triggered by dynamic effects associated with fiber fractures in adjacent fragments or, in some cases, by large strains in the vicinity of the debond. The average initial size of the debonded regions was found to first increase and then decrease with applied strain. In addition, tests with bare fibers and with coated fibers indicate that the average size of debonded regions at saturation increases as the calculated interfacial shear strength decreases, as expected. However, changes in the interfacial shear strength resulting from changes in testing protocol did not correlate with the average debond size at saturation.
Citation: Polymer Composites
Pub Type: Journals
E-glass fibeers, epoxy resin, fiber fracture, interfacial debonding, non-linear viscoelasticity, single fiber fragmentation test, stress concentration