Atomic Force Microscopy Simulations for CO-functionalized tips with Deep Learning

Jaime Carracedo-Cosme,Rubén Pérez

DOI: https://doi.org/10.48550/arXiv.2212.13809

2022-12-28

Abstract:Atomic Force Microscopy (AFM) operating in the frequency modulation mode with a metal tip functionalized with a CO molecule images the internal structure of molecules with an unprecedented resolution. The interpretation of these images is often difficult, making the support of theoretical simulations important. Current simulation methods, particularly the most accurate ones, require expertise and resources to perform ab initio calculations for the necessary inputs (i.e charge density and electrostatic potential of the molecule). Here, we propose an efficient and simple alternative to simulate these AFM images based on a Conditional Generative Adversarial Network (CGAN), that avoids all force calculations, and uses as the only input a 2D ball--and--stick depiction of the molecule. We discuss the performance of the model when optimized using different training subsets. Our CGAN reproduces accurately the intramolecular contrast observed in the simulated images for quasi--planar molecules, but has significant limitations for molecules with a significant internal torsion, due to the strictly 2D character of the input.

Materials Science,Disordered Systems and Neural Networks

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the difficulty in the simulation and interpretation of high - resolution atomic force microscopy (HR - AFM) images. Specifically, an atomic force microscope operating in the frequency - modulation mode with a metal probe functionalized by CO molecules can image the internal structure of molecules with an unprecedented resolution. However, the interpretation of these images is often very difficult and requires the support of theoretical simulations. Current simulation methods, especially the most accurate ones, need to perform ab initio calculations to obtain necessary inputs (such as charge density and electrostatic potential of the sample), which are not only time - consuming but also require specialized knowledge. For this reason, the paper proposes a new method based on conditional generative adversarial networks (CGAN) for efficiently and simply simulating these AFM images. This method avoids all force calculations and only requires a two - dimensional ball - and - stick model of a molecule as input. In this way, the researchers hope to provide a fast and easy - to - use tool to support the understanding and analysis of AFM images of complex molecular structures, especially when the molecules have significant internal distortion and the existing methods have obvious limitations. In addition, this study also explores the influence of different training subsets on the performance of the model, including the selection of the maximum internal undulation and amplitude parameters of the molecules, thus providing directions and suggestions for future research.

Atomic Force Microscopy Simulations for CO-functionalized tips with Deep Learning

Molecular identification with atomic force microscopy and conditional generative adversarial networks

Molecular Dynamics Simulation of Form Measurement Process of Soft Materials Using Atomic Force Microscope

Simulation atomic force microscopy for atomic reconstruction of biomolecular structures from resolution-limited experimental images

Molecular Identification via Molecular Fingerprint extraction from Atomic Force Microscopy images

A Deep Learning Approach for Molecular Classification Based on AFM Images

Automated Structure Discovery in Atomic Force Microscopy

Exploring Molecules Beyond CO As Tip Functionalizations in High-Resolution Noncontact Atomic Force Microscopy: A First Principles Approach

Simulations of Subnanometer Scale Image Contrast in Atomic Force Microscopy of Self-Assembled Monolayers in Water

The mechanism of high-resolution STM/AFM imaging with functionalized tips

Adaptive semi-empirical model for non-contact atomic force microscopy

Simulation of CNT-AFM tip based on finite element analysis for targeted probe of the biological cell

Finite Element Modelling of Atomic Force Microscopy Imaging on Deformable Surfaces

Molecular Identification from AFM images using the IUPAC Nomenclature and Attribute Multimodal Recurrent Neural Networks

Imaging of atomic orbitals with the Atomic Force Microscope - experiments and simulations

Reducing Molecular Simulation Time for AFM Images Based on Super-Resolution Methods

Investigation of the Interaction Between Carbon Nanotube Tip and Silicon Sample Through Molecular Dynamic Simulation

Structural and functional imaging with carbon nanotube AFM probes

Insights and guidelines to interpret forces and deformations at the nanoscale by using a tapping mode AFM simulator: dForce 2.0

Atomic force microscopy as nanorobot.

General Resolution Enhancement Method in Atomic Force Microscopy (AFM) Using Deep Learning