Novel Tools for Studying Crystallization Pathways of Shaped Particles into Complex Crystals

Nucleation and crystallization processes are critical in diverse materials, such as colloidal, soft, molecular, metallic, and ionic crystals. While there's considerable understanding of nucleation in simple crystals, research on complex crystals' nucleation is limited due to the lack of suitable computational tools. We address three main areas for improvement: an order parameter that describes local structures with continuous point group symmetry, a new neighbor list for shaped particles, and a tool for efficiently detecting and recording nucleation events in great detail.<br/>Complex crystals exhibit multiple local environments, and to truly understand the nucleation process on a lowest level would require us to somehow track the progression of symmetry elements for each of these environments. However, traditional order parameters such as Steinhardt's (SOP) and its variants merge multiple symmetries into a single value, limiting their effectiveness in detailed analysis. Our novel Point Group Order Parameter (PGOP) overcomes this limitation by providing a per-particle order parameter based on symmetry operations. PGOP compares a particle's bond order diagram (BOD) with a perfect symmetrized BOD using the Pearson correlation coefficient. This quantifies how well the local environment aligns with specific symmetry operations, allowing distinct differentiation of local environments in complex crystals.<br/>The accuracy of PGOP depends on identifying neighboring particles. Traditional methods like Voronoi tessellation and Solid-angle nearest neighbors (SANN) often fail in systems with shaped particles. To resolve this, we introduce Shadow Projection Overlap of Obstacles for Neighbor Exclusion (SPOONE). SPOONE identifies neighbors by casting rays and determining intersections (blocking), providing more accurate neighbor lists for shaped particles and enhancing PGOP's effectiveness. SPOONE can be utilized as a traditional or weighted neighbor list by considering the amount of blockage (shadow). This makes it very suitable in use for order parameter computation.<br/>To facilitate the automatic detection of nucleation events and manage the data-intensive nature of these simulations, we developed Dupin, a Python tool for event detection in molecular simulation trajectories. Dupin promises to automate detection of events offering new potential applications of machine learning in the field. Dupin identifies changes using order parameters and employs change point detection methods to record events with high temporal resolution. Since dupin can work on-the-fly during the simulation it can also serve to minimizes data storage and improves simulation efficiency by selectively recording only significant events.<br/>To showcase the utility of our newly developed tools, we present results from comparisons in traditional situations suitable for each tool. We contrast PGOP with SOP in both simple and complex crystals. SPOONE's effectiveness is demonstrated in systems of elongated shaped particles, with performance comparably effective for nearly spherical particles. Additionally, we explore the impact of different neighbor lists on PGOP and SOP by comparing the results using SPOONE versus Voronoi methods. Additionally, we highlight Dupin's ability to efficiently identify various events in soft matter simulations, including intricate nucleation processes and polymer collapses, for comprehensive post-simulation analysis. These tools collectively promise to advance the study of nucleation and crystallization, potentially opening new research opportunities in large-scale crystallization projects and enabling application of emerging fields such as machine learning in studying nucleation and crystallization pathways.