A Novel Method for High-Resolution Material Identification via Atomic Force Microscopy Ringing Mode and Machine Learning

We present a novel methodology for high-resolution identification of material composition on sample surfaces utilizing atomic force microscopy (AFM) operating in sub-resonance tapping Ringing mode. This advanced technique leverages the unique capability of Ringing mode to simultaneously acquire multiple physical and mechanical property maps with subnanometer lateral resolution. Material identification is achieved by comparing these high-resolution maps against a database of known material properties. The material recognition is done at each pixel of the AFM image with the help of machine learning algorithms. The efficacy of this approach is demonstrated through its application to blends of distinct polymers, specifically polystyrene (PS), polyvinyl pyrrolidone (PVP), and polyethylene oxide (PEO). By precisely localizing the spatial distribution of these constituent polymers within the sample, our methodology enables detailed characterization of complex polymer systems with unprecedented resolution. Furthermore, we provide a comparative analysis of the advantages and limitations of our Ringing mode AFM and machine learning-based technique with respect to other established spectroscopy methods, such as confocal Raman and AFM-IR microscopy. This in-depth evaluation offers valuable insights into the potential applications and future developments of this cutting-edge approach in the field of material characterization and beyond.