Advancing Scanning Probe Microscopy Simulations: A Decade of Development in Probe-Particle Models

The probe-particle model is an open-source package designed for simulation of scanning probe microscopy experiments, employing non-reactive, flexible tip apices (e.g., carbon monoxide, xenon, or hydrogen molecules) to achieve sub-molecular resolution. This abstract introduces the latest version of the probe-particle model, highlighting substantial advancements in accuracy, computational performance, and user-friendliness over previous versions. To demonstrate this we provide a comprehensive review of theories for simulating non-contact Atomic Force Microscopy (nc-AFM), spanning from the simple Lennard-Jones potential to the latest full density-based model. Implementation of these theories are systematically compared against ab initio calculated reference, showcasing their respective merits. All parts of the probe-particle model have undergone acceleration by 1-2 orders of magnitude through parallelization by OpenMP on CPU and OpenCL on GPU. The updated package includes an interactive graphical user interface (GUI) and seamless integration into the Python ecosystem via pip, facilitating advanced scripting and interoperability with other software. This adaptability positions the probe-particle model 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 of 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 surface science and molecular chemistry.