Kousika

Understanding the influence of defects on interfacial properties of hexagonal boron nitride – water system using deep neural networks

2D materials like graphene, hexagonal boron nitride (hBN) and molybdenum disulphide (MoS2) are emerging as potential materials for water desalination because of their high surface-to-volume ratio and high-water permeability due to atomic thinness. Among them, hBN and graphene have comparable water permeability and ion selectivity. In addition, hBN has high mechanical and chemical stability (especially towards oxidation) than graphene and graphene-based materials making it more suitable for membranes. As the properties of water in nanoconfinement are different from bulk water, it is important to understand water- 2D material interactions at the atomic-level. Classical and ab-initio based molecular dynamics (MD) simulations are widely used to understand the water permeation through the nanochannels of 2D materials. MD gives valuable information on the rates of transport of molecules and ions through nanochannels, helping understand water permeation and salt rejection. Though studies using classical MD potentials give useful molecular insights, they are mostly empirical which affects their accuracy and transferability.

To overcome this limitation, machine-learning (ML) potentials are being explored to describe the interaction of water-2D material systems with the accuracy of ab-initio methods and the efficiency of classical MD. Such ML potentials are derived for pristine hBN- and pristine graphene-water systems using neural networks to understand the water permeation through nanocapillaries. Given the importance of accurately modelling realistic water-hBN systems, in this work, ML potentials are developed for water – defective hBN system using density functional theory calculations. After that, these potentials are used to understand the influence of various point defects on the interfacial properties such as friction coefficient and slip length of hBN-water system.