Application of artificial neural networks in atomic force microscopy

Abstract

The paper describes a method of artificial neural network application to build a computer database for numerical simulation of the process of indentation of an atomic force microscope probe into an elastomeric composite with a granular filler (nonlinear elastic medium with rigid spherical inclusions). The use of such a base makes it possible to significantly improve the speed and quality of interpretation of the results of nanoindentation for structurally inhomogeneous materials. In this case, information becomes available not only about what is done on the surface of the sample but also in the near-surface layer inside it. An algorithm has been developed with the help of which an artificial neural network was built and “trained,” designed to obtain indentation curves depending on the size of the filler particles and its localization in the near-surface layer of the composite (depth and horizontal distance from the top of the AFM probe). It is shown that the speed of constructing indentation curves increases by several orders of magnitude compared to conventional approaches based on the numerical solution of the corresponding boundary value problems for each specific case. Accordingly, computer costs are also significantly reduced, that is, in the presence of an already built and “trained” neural network, powerful and high-speed computers are not needed.