On failure phenomenon in single-fibre pull-out tests

Abstract

This letter seeks to provide a theoretical explanation of a commonly observed failure phenomenon in the single-fibre pull-out test, namely the presence of a matrix cone at the emergent end of the extracted fibre. The single fibre pull-out test is commonly used [1,2] to characterize fibre reinforced polymer composites. This is because the test is simple to conduct and can be Used for almost any fibre-matrix composite system. An advantage [1] of the pull-out test is that in addition to the interfacial bond strength and interfacial toughness, other interfacial properties, such as the matrix shrinkage pressure on the fibre, the interfacial shear stress and the work done in pulling the fibre from the matrix, can be determined. The last factor is important, since the significant increase in fracture toughness of fibrous composites has been attributed to the fibre pull-out process during failure. Four types of failure [3] may occur in the specimen during a single-fibre pull-out test. These are: (i) specimen failure due to matrix failure away from the fibre-matrix interface, (ii) specimen failure by fibre fracture along the external free length of the fibre, (iii) partial debonding followed by specimen failure due to fibre fracture within the embedded length of the fibre and (iv) specimen failure due to complete debonding and extraction of the debonded fibre from the matrix. The first three types of failure represent unsuccessful pull-out tests. Data from specimens that exhibit these types of failure should not be included in pull-out analysis. Only specimens that exhibit failure of type (iv) represent successful pull-out tests. Thus, only data from such specimens should be utilized for analysis. This letter concerns specimens that exhibit this type of failure. A matrix "cone" is frequently observed at the emergent end of the extracted fibre after specimen failure. Such a typical matrix cone on the extracted fibre of a glass fibre/epoxy matrix pull-out specimen is shown in Fig. l a. The region below the matrix cone is the exposed debonded fibre surface. The point at which the cone is widest represents the original location of the edge of the matrix block that was bonded to the fibre. The part of the matrix cone above this location represents the menscus formed by the epoxy on the glass fibre. The matrix cone of an extracted fibre from a specimen with short embedded length is shown in Fig. lb. All of the features described above for a specimen with larger embedded length can be seen clearly in Fig. lb. Such matrix cones have been reported in the literature for both thermoset-based [3, 4] and thermoplastic-based [5] single-fibre pull-out specimens. None of the existing models [6-8] can account for the formation of such a matrix cone during type (iv) failure of a single-fibre pull-out specimen. It will be shown in this letter that radial and circumferential stresses in the matrix are responsible for the formation of the matrix cone. Radial stresses O'rm are usually neglected in the existing models, which predict a maximum finite value of the interfacial shear stress ~'i at the fibre emergent end since oS~ is much smaller [6] "ri. Therefore O~m is expected to have little effect on debonding crack initiation at the emergent end of the single-fibre pull-out specimen. However, the prediction of a finite maximum