Studies of fiber-matrix adhesion in composites using fractal theory, SAXS and SEM

The effect of O₂ plasma etching on the morphology of graphite fibers was studied using small angle x-ray scattering (SAXS), scanning electron microscopy (SEM) and fractal theory. These tools were used to study the hypothesis that plasma etching produces changes in the surface topography and chemical reactivity of the fiber surface, and these changes would affect the adhesion characteristics between the fibers and the matrix. In order to determine the effect of etching on the mechanical properties of the fibers of the composites made from both treated and untreated fibers, mechanical testing was done using an Instron Mechanical Tester.

The void volume fraction and scattering invariant of both pitch and polyacrylonitrile-based graphite fibers were shown by SAXS to be greatly affected by plasma etching. For example, the measured volume fraction of voids in the high modulus PAN-based fibers was 2.92% for the unetched case, reached a minimum of 0.50% at one—half minute etch time, and gradually increased to a value of 1.46% at four minutes etching. The SAXS results indicated that the plasma preferentially etches the larger voids first and then smaller voids are produced in the fiber as the etching process continues. SEM micrographs showed a marked roughening of the fiber surface and the development of etch pits as a result of the plasma etching process.

The mechanical properties of the fibers themselves were not seriously degraded as a result of etching. The interlaminar shear strength (ILSS) of the composites was measured for each fiber treatment time as a measure of fiber-matrix adhesion. The increased roughness of the fiber surface as well as the possible incorporation of reactive chemical groups onto the surface as a result of O₂ plasma etching should produce ILSS values that increase with increasing etching time. However, the ILSS of the etched fiber composites did not exhibit a dramatic increase as the etch time increased. The difference in ILSS values from untreated to four minutes etch time was about 10%, which may be attributed to experimental uncertainty. This was probably due in part to the low volume fraction of fibers (an average of 11%) that was incorporated into the composites.

Since the ILSS did not change greatly with etch time and due to the fact that the fractal dimension did not seem to change with etch time (again about 10% between untreated and four minutes etch time), the attempt to correlate fractal dimension with adhesion characteristics was only marginally successful. However, SEM micrographs of the failure surface of the composites showed a marked difference between etched and unetched fibers. As the etching time increased, the degree of fiber pullout decreased and also the length of broken fibers extending from the failure surface decreased. Thus, the SEM studies indicated that while the ILSS results did not conclusively show the effect of increased adhesion, plasma etching did improve fiber-matrix adhesion, at least on a local scale.