A surface evolution scheme to identify nanoscale intrinsic geometry from AFM experimental data

The geometrical properties of metallic nanoparticles such as the size and morphology have significant impacts on the structure and stability of the adsorbed biological entities as well as the nanoscale structural performances. To identify the nanoscale intrinsic geometry from the height images by atomic force microscopy (AFM), we developed a curvature-dependent evolution scheme that can eliminate the noise and smoothen the surfaces. The principal curvatures are computed directly from the first and second derivatives of the discrete AFM height data. The principal curvatures and directions correspond to the eigenvalues and eigenvectors of shape operator matrix, respectively. The evolution equation using the principal curvature flows smoothens the images in the corresponding principal directions. For an idealized model, κ 2 flow successfully identifies the major valley lines to represent the boundary of nanoparticles without referring to the phase information, whereas the mean curvature flow eliminates all the minor ones leaving only the major feature of the boundary. To demonstrate the capability of noise removal, smoothing surfaces, the identification of ridge and valley lines, and the extraction of intrinsic geometry, the developed numerical scheme is applied to real AFM data that include the silver nanoparticles of 24 nm diameter and the gold nanoparticles of 33–56 nm diameters.