Spectral analysis of irregular roughness artifacts measured by atomic force microscopy and laser scanning microscopy

Atomic force microscopy (AFM) and laser scanning microscopy (LSM) measurements on a series of specially designed roughness artifacts were performed and the results characterized by spectral analysis. As demonstrated by comparisons, both AFM and LSM can image the complex structures with high resolution and fidelity. When the surface autocorrelation length increases from 200 to 500 nm, the cumulative power spectral density spectra of the design, AFM and LSM data reach a better agreement with each other. The critical wavelength of AFM characterization is smaller than that of LSM, and the gap between the measured and designed critical wavelengths is reduced with an increase in the surface autocorrelation length. Topography measurements of surfaces with a near zero or negatively skewed height distribution were determined to be accurate. However, obvious discrepancies were found for surfaces with a positive skewness owing to more severe dilations of either the solid tip of the AFM or the laser tip of the LSM. Further surface parameter evaluation and template matching analysis verified that the main distortions in AFM measurements are tip dilations while those in LSM are generally larger and more complex.