Eg] EH HE Ens

Figure 2 shows a representative sample of indents near a fiber with the corresponding indentation depths. Nanoindentation was performed on the 0%, 0.1% and 1.0% APS samples; however, the indents made into the area where the measured interphase was located contained fiber bias effects. This was most apparent in the 0% and 0.1% samples. For example, the 0% APS sample did not yield a clearly evident interphase; therefore, the indents made near the fiber with a 16 /xN load should have had the same penetration depth as those in the matrix. However, as the indents progressed near the fiber, this penetration depth decreased. This phenomenon is also apparent in the 0.1% and 1.0% APS samples that had a measurable interphase. The percent difference between the indentation depths in the interphase and matrix

of the 0.1% sample is —7.6%. Ultimately, this signifies that the softer interphase, as viewed by phase imaging, was manifesting as a stiffer body when mechanically analyzed by nanoindentation. This conclusion contradicted what was found previously through phase imaging and is most likely due to fiber bias effects. The 1.0% APS sample yielded a percent difference of 6.6%. This positive value signifies that the interphase indents had a deeper penetration depth than those made into the matrix, and agrees with the qualitative measurements from AFM-PI . However, the interphase indents may have been compromised by fiber bias effects, similar to the effect found for the 0% and 0.1% APS samples. Also, prior work in our research group [25] showed that, for nanoindentation near glass fibers, there was a linear increase in the indentation modulus as a function of distance from the fiber surface. Extrapolating Kumar's data [25] to 16 /xN indicates that fiber bias effects would decrease the indentation depth approximately 53% at 110 nm, 46% at 210 nm, 33% decrease for 375 nm and 17% decrease for 888 nm.