An AI-Driven Framework for Generating Optimal Structures in Material Science Using Generative Methods

Research question

In the field of sustainable materials science, understanding and manipulating nanomaterial growth remains a daunting task. Traditional atomic scale methods such as ab initio molecular dynamics (aiMD) based on density functional theory (DFT) are often too computationally expensive to access the nano to millisecond time scale needed to study growth mechanisms. To overcome these limitations, we propose an AI-driven framework that simplifies the construction of machine learning force fields (MLFFs) trained on DFT results. Our recent work on developing a MLFF (DeepCNT-22) capable of modeling single-wall carbon nanotube growth provided new insights into the formation of interface defects and the conditions needed for defect-free growth. Aiming for further enhancement, our research seeks to expand DeepCNT-22 to incorporate hydrogen, targeting a detailed understanding of hydrocarbon feedstock gas decomposition in material growth processes. With the goal of creating an atomic-level digital twin of CVD/ALD material growth. However, incorporating hydrogen is accompanied by the challenge of efficiently selecting and generating representative atomic configurations. Our novel idea is to develop an AI-driven framework that uses advanced machine learning models and physics-informed learning to optimize atomic environments descriptors and to use this framework to select and generate representative atomic configurations for MLFF training. This development not only promises significant efficiencies in MLFF construction, reducing reliance on expensive DFT calculations, but also lays the groundwork for more extensive future explorations in materials science. A strong synergy between WISE and WASP.

Sustainability aspects

This framework addresses the challenges faced by MLFFs in adapting to varied chemical environments. Many existing models struggle to generalize across different systems without retraining, which is time-intensive and resource-consuming. By focusing on robust, quality dataset generation, our framework optimizes MLFF performance and resilience. In collaboration with WISE and WASP, we are poised to make impactful advances in nanomaterial growth control, enabling applications in energy-efficient carbon nanomaterials for electronics and energy storage. This proposed framework will play a crucial role in realizing defect-free carbon nanomaterials with consistent properties, advancing their use in sustainable technology solutions. The scalability and efficiency of these MLFF models will set new benchmarks for computational materials research, opening pathways to address global sustainability challenges and fostering innovations in clean energy and green technologies.