The surface roughness is a key parameter for the characterization of surfaces with respect to interfacial properties. The surface roughness of a carbon fiber affects both the fiber-matrix bonding and the composite failure behavior, due to its influence on effective surface area and mechanical interlocking between fiber and matrix.
The topography of a surface can be formed by structures on different length scales. It can be measured by Atomic Force Microscopy (AFM) with great accuracy.
One of the challenges of the roughness characterization is the data reduction. In a traditional approach, surface roughness parameters are evaluated based on averaging of absolute height values. However, these averages are not solely determined by the extensions of these structures but can be dominated by large height differences.
This drawback is addressed in a previous publication of our group. This approach is based on a background correction that eliminates surface structures larger than a chosen length scale. However, only surface structures smaller than the chosen length scale are considered.
Here, we present a new approach to quantify surface roughness on all length scales covered by AFM measurements. The method applied uses Fourier transformation to calculate the Power Spectral Density (PSD), which describes the contribution of different spacial frequencies to the surface roughness and whose integral equals the root mean square roughness according to Parseval’s theorem. Thus, different surfaces can be compared intuitively by graphical representations of the PSD and roughness as a function of the extent of the respective structures. To demonstrate this method, commercially available PAN-based carbon fibers of different types and exposed to different surface treatments are measured by AFM.
It is shown that surface roughness quantification by PSD is a convenient method to evaluate and visualize the contributions of structures of different extent to the overall surface roughness.