This study employs Molecular Dynamics simulations to enhance the structural characterization of sphere mixture systems, focusing on local order parameters to identify order-disorder transitions within lamellar and hexagonal mesophases. The structural characterization included the use of neural networks and the search for clusters with topological order. We have identified the order-disorder transition at a characteristic temperature TOD, effectively captured by the structural parameters analyzed. Below TOD, the system organizes into an ordered mesophase, while above this temperature, it transitions into an isotropic liquid state. Moreover, within the isotropic region, we observe that the liquid begins to form clusters. This clustering phenomenon becomes evident below a specific temperature, denoted as T⁎. At T⁎, there are noticeable changes in the volume versus temperature curves, indicating a process of micellization or clustering in the system. The structural parameters used in our study also register this behavioral change at T⁎, further supporting the occurrence of this phenomenon. Additionally, our analysis of the overlap in detecting ordered particles between the two structural parameters using the Sørensen-Dice similarity coefficient reveals a significant spatial correlation in the isotropic phase, with a correlation higher than what would be expected in a random case. This indicates that both neural networks and topological cluster classification detect statistically similar ordered regions in the isotropic phase, highlighting the clustering. Our findings reveal that the structural parameters effectively capture both the order-disorder transition at TOD and the clustering process at T⁎. These results provide valuable insights into the self-assembly mechanisms and phase behavior in this soft-sphere model, advancing our understanding of self-assembly processes in material science, particularly in exploring phase behavior in mesogenic nanoparticle systems.
Read full abstract