Phan, Tien Minh2021-07-302021-07-30https://hdl.handle.net/2097/41585The study of molecular self-assembly has attracted considerable interest over the decades due to its wide-ranging applications in chemistry, materials science, and biology. The challenge in studying molecular self-assembly is that it involves complex spatiotemporal scales ranging from short-lived microscopic events to lifelong macroscopic architectures. This work aims to investigate the molecular self-assembly in different systems to better understand the macroscopic properties of microscopic activities. To overcome the complexity of time scale and length scale, we employed a combination of computer simulations and simple analytic theories to develop multi-scale models to study the systems of interest. We find that: (1) in the slow growth regime of crystalline solids, impurity particles can speed up crystal growth with minimal impact on the final product, (2) in the self-assembly of peptides into amyloid fibrils the conformational entropy plays an important role in the transition from nucleation to elongation, (3) in biomolecule condensates, the surface tension arises from the competition between binding energy and configurational entropy. These results highlight the power of multi-scale models to interpret macroscopic physical observables in terms of fundamental microscopic mechanisms in molecular self-assembly processes.en-USMulti-scale modelsSelf-assemblyProtein aggregationClassical nucleation theoryMonte Carlo simulationsMD simulationsDissertation