Characterizing the fiber-matrix interphase via single fiber composite tests.

摘要

The assumptions of weakest-link theory and statistics were considered, especially as they relate to high-strength and high-modulus carbon fibers. A critical review of proposed failure mechanisms and flaw populations was conducted. Numerical simulations and morphological evidence showed that the Reynolds-Sharp failure criteria for transverse rupture of misoriented turbostratic graphite crystallites are infrequently satisfied. Furthermore, the mean crystallite size in these carbon fibers is too small to cause catastrophic tensile failure, except in cooperation with other misoriented crystallites. A Reynolds-Sharp failure-strength model and a phenomenological/statistical model developed by Batdorf and Crose were used to predict failure statistics for nonuniformly oriented flaw populations exposed to polyaxial stress states.; Ultimate tensile tests were conducted on carbon fibers of various gauge lengths and surface treatments. Results indicate two distinct flaw populations in these PAN-based carbon fibers and that their failure statistics are represented by a concurrent Weibull distribution with a broad, lower-strength surface flaw population, and a tighter, higher-strength volume distribution.; Next, we explored the morphology, chemistry, and micromechanics of the fiber-matrix interphase and their influence on fiber-matrix load transfer. In particular, the development, utility, accuracy, and analytical interpretation methods for the single fiber composite fragmentation (SFCF) test are investigated. The complex nonlinear three-dimensional stress state near fiber-matrix imperfections in conjunction with irregular interphase properties discounts simple shear-lag or yielded-matrix micromechanical analysis. It is shown, however, that interpretation approaches that include fiber-matrix debonding and fiber strength variation, while more representative of the physical fragmentation process, are unable to accurately represent the stochastic fragmentation process. Additionally, it is argued that the interfacial shear strength microvariable tau ifss does not unambiguously define matrix-fiber load absorption.; A non-deterministic model and analysis (NDA) was developed for simulating the SFCF test and its attendant uncertainties, reducible and non-reducible. The NDA demonstrated that the accumulated uncertainties in SFCF testing, modeling, and interpretation contribute to a large response prediction uncertainty for tau ifss. The utility and suitability of SFCF testing is therefore questioned, especially if the results of the tests are reported without confidence intervals.