Browsing by Author "Hapala, Prokop"

Now showing 1 - 3 of 3
  • Thumbnail Image
    Item
    Advancing scanning probe microscopy simulations : A decade of development in probe-particle models
    (Elsevier Science B.V., 2024-12) Oinonen, Niko; Yakutovich, Aliaksandr V.; Gallardo, Aurelio; Ondráček, Martin; Hapala, Prokop; Krejčí, Ondřej; Department of Applied Physics; Surfaces and Interfaces at the Nanoscale; Computational Electronic Structure Theory; Swiss Federal Laboratories for Materials Science and Technology; Fundacion IMDEA Nanociencia; Czech Academy of Sciences
    The Probe-Particle Model combine theories designed for the simulation of scanning probe microscopy experiments, employing non-reactive, flexible tip apices to achieve sub-molecular resolution. In the article we present the latest version of the Probe-Particle Model implemented in the open-source ppafm package, highlighting substantial advancements in accuracy, computational performance, and user-friendliness. To demonstrate this we provide a comprehensive review of approaches for simulating non-contact Atomic Force Microscopy. They vary in complexity from simple Lennard-Jones potential to the latest full density-based model. We compared those approaches with ab initio calculated references, showcasing their respective merits. All parts of the ppafm package have undergone acceleration by 1-2 orders of magnitude using OpenMP and OpenCL technologies. The updated package includes an interactive graphical user interface and seamless integration into the Python ecosystem via pip, facilitating advanced scripting and interoperability with other software. This adaptability positions ppafm as an ideal tool for high-throughput applications, including the training of machine learning models for the automatic recovery of atomic structures from nc-AFM measurements. We envision significant potential for this application in future single-molecule analysis, synthesis, and advancements in surface science in general. Additionally, we discuss simulations of other sub-molecular scanning-probe imaging techniques, such as bond-resolved scanning tunneling microscopy and kelvin probe force microscopy, all built on the robust foundation of the Probe-Particle Model. Altogether this demonstrates the broad impact of the model across diverse domains of on-surface science and molecular chemistry.
  • Thumbnail Image
    Item
    Interpreting atomic force microscope images with machine learning
    (2019-08-20) Oinonen, Niko; Hapala, Prokop; Urtev, Fedor; Perustieteiden korkeakoulu; Foster, Adam
    Since its invention in 1986, atomic force microscopy (AFM) has developed into a unique tool for exploring the microscopic world. With the introduction of CO tip functionalization, the image resolution has reached the level of individual atoms and bonds. However, the use of this method so far has been mostly restricted to planar structures, due to difficulties in interpretation of images for more complex 3D molecular structures. We aim to address this problem with the use of artificial neural networks (ANN), a type of machine learning model. ANNs have gained much attention in recent years for advancing the state of the art in many complex problems, including those related to natural language processing, image recognition, autonomous cars, and playing games at a superhuman level. The success of ANNs has been enabled by the increased availability of computational resources and datasets of sufficient size. In the work of this thesis, we apply convolutional neural networks, a type of ANN, to the task of predicting easily interpretable descriptors of atomic properties from AFM images. The models are trained on simulated AFM images and tested on both simulated and experimental images. The results on simulated images are generally very good, but experimental results, while in some cases promising, indicate that there are some challenges that need to be overcome.
  • Thumbnail Image
    Item
    Quantifying the evolution of atomic interaction of a complex surface with a functionalized atomic force microscopy tip
    (Nature Publishing Group, 2020-12-01) Liebig, Alexander; Hapala, Prokop; Weymouth, Alfred J.; Giessibl, Franz J.; Department of Applied Physics; University of Regensburg
    Terminating the tip of an atomic force microscope with a CO molecule allows data to be acquired with a well-known and inert apex. Previous studies have shown conflicting results regarding the electrostatic interaction, indicating in some cases that the negative charge at the apex of the CO dominates, whereas in other cases the positive charge at the end of the metal tip dominates. To clarify this, we investigated CaF 2(111). CaF 2 is an ionic crystal and the (111) surface does not possess charge inversion symmetry. Far from the surface, the interaction is dominated by electrostatics via the negative charge at the apex. Closer to the surface, Pauli repulsion and CO bending dominate, which leads to an unexpected appearance of the complex 3-atom unit cell. We compare simulated data in which the electrostatics are modeled by point particles versus a charge density calculated by DFT. We also compare modeling Pauli repulsion via individual Lennard–Jones potentials versus a total charge density overlap. In doing so, we determine forcefield parameters useful for future investigations of biochemical processes.