(154b) Low-Dimensional Fluctuating Interfaces for Efficient Separations

Fluctuating, low-dimensional soft and hard material interfaces involve physical phenomena at various length scales. For example, the electronic structure of the low-dimensional material has a strong influence on the organization of the soft matter near the material. Atomic and molecular scale description of the materials is important to accurately describe the structure, dynamics, and thermodynamics of materials. Many nanomaterial applications involve macroscopic length scales that cannot be realistically modeled by atomic and molecular scale approaches. As a result, multiscale approaches linking various length scales are essential for many applications involving soft and hard nanomaterials. In this talk, we will present data-driven, machine learning-based multiscale approaches to model soft/hard nanomaterials. First, we will describe a deep learning-based multiscale approach to link ab-initio and molecular dynamics methods. We will show that the development of such methods can provide physical insights into nanomaterials that are not possible with purely ab-initio or molecular approaches. Second, we will describe deep learning-based multiscale approaches to coarse-grain materials at larger length scales. These methods can model systems at larger length scales and provide physical insights that are not possible with purely molecular approaches. Finally, using the data-driven multiscale approaches, we investigate dynamic interfaces involving low-dimensional materials and show that vibrational coupling has a significant effect on interfacial properties such as wetting, friction and fluid slip. We show how vibrational coupling can be tuned to design materials for efficient separations.