Parts | Days | Selection | Search | Updates | Downloads | Help
O: Fachverband Oberflächenphysik
O 32: Poster: Solid-Liquid Interfaces
O 32.14: Poster
Tuesday, March 19, 2024, 18:00–20:00, Poster C
Tip Classification of High Resolution AFM Imaging in Liquids — •Farzin Irandoost1, Fillippo Federici Canova2, Takeshi Fukuma3, Tobias Dickbreder4, Franziska Sabath4, Ralf Bechstein4, Angelika Kühnle4, and Adam S. Foster1,3 — 1Department of Applied Physics, Aalto University; Helsinki, Finland — 2Nanolayers Research Computing Ltd., London, England — 3Nano Life Science Institute (WPI-NanoLSI), Kanazawa University; Kanazawa, Japan — 4Physical Chemistry I, Bielefeld University, Germany
AFM imaging in liquids is profoundly influenced by scanning height and tip, leading to different 2D maps for the same crystal surface. This study introduces a workflow to discern diverse contrast patterns arising from AFM scanning height and tip dependency. Our workflow, firstly, clusters images in large experimental datasets based on contrast pattern similarities, then links the clusters to relevant free energy simulations based on the tip characteristics.
Success relies on selecting a sensitive latent vector and implementing a robust clustering benchmark. While statistical analysis of Fourier transforms peaks as image descriptors, aided by tSNE and K-means clustering shows promise in qualitative evaluation, developing a quantitative method for large dataset evaluation remains a priority. To this end, we are developing a method for lattice vector extraction out of experimental data, which makes quantitative benchmarks available for periodic crystal images. Also, the latter can be used as an alternative way of describing images containing periodic patterns.
Keywords: Solid-Liquid Interface; Atomic Force Microscopy; Unsupervised Learning