<p dir="ltr">Computational materials science has emerged as a powerful technique to discover and develop new materials in past decades, primarily because accurate computational modeling can act as guidance before performing experiments that are expensive and time-consuming. However, modeling material behaviors across different scales of length and time poses a challenge, accentuating the importance of choosing appropriate levels of approximations and theories. First principles calculations based on density functional theory (DFT) are essential to predict the electronic structure of periodic crystalline systems. We will discuss a prediction of chemical doping induced metal-to-insulator transition (MIT) of transition metal perovskites owing to the variation of the electronic occupation. Nevertheless, electronic structure predictions based on DFT are not without limitation as it fails when treating strongly correlated electronic system due to the over-delocalization of valence electrons. In principle, adding on-site Hubbard U corrects this error with a low computational cost. Using an example of a two-dimensional rare-earth MXene, we demonstrate the essence of choosing the appropriate U value self-consistently for the prediction of electronic and magnetic configurations. Furthermore, molecular dynamics (MD) can be employed to study the dynamic evolution of complex condensed systems with thousands to millions of atoms at the atomistic and molecular levels. Carbon fiber manufacturing is an established industry, though the fiber produced achieves only 10% of its theoretical tensile strength. Therefore, optimizing the carbon fiber processing is a pressing topic. To achieve this, we study two steps, spinning and stabilization, of polyacrylonitrile (PAN)-based fiber fabrication at the molecular level using MD. We will discuss the realistic molecular structure of the spun PAN and the properties affected by its structural heterogeneity. Moreover, for the following step, we develop a PAN stabilization simulator, an automated workflow that addresses the underlying chemistry and the molecular-level structure-property relationship, often inaccessible through experiments.</p>
| Identifer | oai:union.ndltd.org:purdue.edu/oai:figshare.com:article/24592995 |
| Date | 20 November 2023 |
| Creators | Shukai Yao (17419314) |
| Source Sets | Purdue University |
| Detected Language | English |
| Type | Text, Thesis |
| Rights | CC BY 4.0 |
| Relation | https://figshare.com/articles/thesis/MULTISCALE_MODELING_OF_POLYMER_PROCESSING_AND_ELECTRONIC_MATERIALS/24592995 |
Page generated in 0.002 seconds