AI-Based Chirality Classification for Carbon Nanotubes

Carbon nanotubes (CNTs) possess exceptional properties such as high conductivity, lightweight, high strength, and excellent heat dissipation, making them ideal for advanced material applications. A significant challenge in CNT research is controlling chirality, which directly influences their electrical properties, distinguishing between metallic and semiconducting CNTs. Traditional chirality classification methods, involving the analysis of diffraction images obtained through Transmission Electron Microscopy (TEM), are complex and labor-intensive. To address this challenge, we developed an AI approach aimed at automating the chirality classification process. The AI was trained on a comprehensive dataset of simulated diffraction images representing various CNT chiralities and tilt angles, ensuring robust training. In tests using experimental TEM images, the AI achieved a classification accuracy of 97%. This high accuracy enables a more efficient classification process. This approach not only enhances the efficiency of CNT research but also opens new avenues for leveraging machine learning in materials science. It addresses complex challenges in materials characterization and advances the broader field of nanotechnology and advanced materials research.