The adhesion of elastomer‐coated glass fibers to an epoxy matrix. Part I: The effect of the surface treatments on the tensile strength of the glass fibers

Abstract

E‐glass fibers were subjected to various surface treatments to study the interfacial adhesion with an epoxy matrix by means of the fragmentation test. The glass fibers considered were both untreated and treated with γ‐aminopropyltriethoxysilane (γ‐APS). In addition, glass fibers were coated with a thin layer of a crosslinkable elastomer including (or not) a silane coupling agent. To evaluate the effect of the coating process, the glass fibers were also passed through the pure solvent (washed fibers). The tensile strength of glass fibers at short length (near the critical length, lc, for the fragmentation test) cannot be measured directly, thus, extrapolation techniques were used. The Weibull statistics technique was applied and accurately described the tensile strength data of the tensile strength at various fiber gauge lengths for the different surface treatments. Nevertheless, the Weibull parameter, m, changes with the gauge length, and therefore, the extrapolation value at lc depends on the method. The tensile strength of the silane‐treated glass fibers is higher than for the untreated fibers, in agreement with reported data. This effect could be attributed to the protection against water provided by the silane sizing. The coating process does not induce any damage of the glass fibers since the washed fibers display mechanical properties close to those obtained for the untreated fibers. An additional effect is observed for the elastomer‐coated fibers which present the highest tensile strength. This effect could be attributed to an improved protection and/or elimination of the weakest filaments in the glass strand during the treatment.

Extrapolation techniques at short gauge lengths based on the weakest link concept for fibres exhibiting multiple failure modes

“…The in¯uence of the surface treatment and coating on the distributions were discussed. The value of the Weibull modulus determined at a single gauge length is di erent from that determined from a ln (strength) versus ln (gauge length) plot (Ahlstrom 1988, Wagner 1989a). Several models only need tests performed at a single gauge length, such as the Weibull formalism.…”

“…The recent success of the fragmentation test has given rise to studies dealing with the mechanical properties of ® bres (Asloun et al 1989, Ahlstrom andGerard 1995). Obtaining information from this type of interfacial tests requires knowing the tensile strength at the critical fragment length, which is for instance, for a common epoxy/ E-glass ® bre interface, in the range 0.3± 0.6 mm (Ahlstrom 1988).…”

Extrapolation techniques at short gauge lengths based on the weakest link concept for fibres exhibiting multiple failure modes

“…The recent success of the fragmentation test has given rise to studies dealing with the mechanical properties of fibres (Asloun et al 1989, Ahlstrom andGerard 1995). Obtaining information from this type of interfacial tests requires knowing the tensile strength at the critical fragment length, which is for instance, for a common epoxy/ E-glass fibre interface, in the range 0.34.6mm (Ahlstrom 1988).…”

“…Strength characteristics of each element are assumed to be equivalent and are generally described in the case of fibres by means of the Weibull (I95 I) statistics. Different approaches have been derived from this formalism, unimodal (Ahlstrom 1988, Ahlstrom andGerard 1995) and bimodal (Beetz 1982, Jung et al 1993) cumulative distribution functions (CDFs), as well as strength-length relationships (Ahlstrom 1988, Asloun 1989. Direct use of the loading history of a singlefibre fragmentation test has also been conducted to calculate the length effect on fibre strength, the Weibull shape and the scale parameters (Yavin et al 1991).…”

Effect of surface treatment on the hydrolytic stability of E-glass fiber bundle tensile strength

“…This type of failure mechanism cannot be accurately described by a single average strength value. While ''weakest link'' methods have been presented in the literature [24][25][26], a different model must be adopted to capture the distribution of strengths found when testing fiber bundles.…”