Oriented attachment (OA) is a significant crystal growth mechanism where nanocrystals attract and attach along specific crystallographic orientations. This process leads to diverse crystal morphologies such as rods, chains, and branched nanowires, and has been observed in various systems including semiconductors, metal oxides, and other materials. However, a quantitative understanding of OA remains elusive, particularly concerning the interplay between crystal and solvent structures, interparticle forces, and ensemble particle dynamics. Previous studies, while providing real-time observations using in situ microscopy techniques, haven't fully elucidated the nature and scaling of the underlying forces or the relationship between nanocrystal structure and surface chemistry, solvent structure, interparticle forces, and particle response dynamics. This lack of comprehensive understanding hinders the rational design of complex materials using OA. This research aims to bridge this knowledge gap by investigating OA of ZnO, a model oxide semiconductor forming quasi-one-dimensional nanostructures with applications in various fields. The choice of ZnO is strategic because it readily forms these nanostructures, and its morphology has been attributed to OA, although direct evidence and detailed dynamic information remain scarce. The study will overcome limitations in traditional liquid-phase transmission electron microscopy (LP-TEM) observations of metal oxides by developing a facile synthesis method and optimizing imaging conditions to mitigate dissolution problems and monitor nanoparticle behavior over long timescales.

Prior research on OA has largely relied on ex situ observations, which offer limited insight into the actual attachment dynamics. In situ microscopy, such as LP-TEM and atomic force microscopy (AFM), has provided more detailed views, but has often focused on specific systems (e.g., iron oxide, Pt-Fe, Au) without a comprehensive analysis of forces involved. Studies have used AFM to measure forces between crystal surfaces, but the relationship between solution composition and structure, and the forces driving OA remains poorly understood. While studies have observed the rotation and alignment of nanoparticles prior to attachment, a full understanding of the underlying forces (e.g., Coulombic, van der Waals, dipole-dipole) and their influence on dynamics has been missing. Existing molecular dynamics (MD) simulations, while helpful, often fall short of predicting the long-range interactions and torques observed experimentally. Therefore, this study aims to improve our understanding by incorporating both experimental and computational techniques to unravel the mechanism of OA in ZnO nanostructures.

This study employs a multi-pronged approach combining in situ LP-TEM observations with computational simulations. First, a facile synthesis method for ZnO nanoparticles was developed to avoid the hydrothermal conditions typically incompatible with LP-TEM, involving the reaction of zinc acetate dihydrate and KOH in methanol at 60°C, followed by growth at room temperature. However, initial LP-TEM imaging revealed rapid ZnO dissolution in pure methanol and water. To address this, the nanoparticles were placed in methanol solutions with increasing concentrations of zinc acetate, ultimately identifying a concentration (1 mM) that stabilized the nanoparticles during imaging. This allowed for extended monitoring of particle behavior under the electron beam. In situ LP-TEM images were then captured to observe the growth of ZnO nanostructures, with movies recorded to track particle trajectories and attachment events. To analyze particle motion, in-house MATLAB algorithms were used to detect and track particles based on image processing techniques, enabling the determination of particle positions, areas, and orientations using Fourier transform analyses. The radial distribution function (RDF) was calculated from the collective behavior of particle ensembles to derive the interparticle potential, W(h). The attachment rate coefficient was calculated by assuming second-order kinetics and a quasi-2D approximation. Mean squared displacements were calculated to determine the diffusion coefficient. The stability ratio was also calculated to compare the experimentally observed coagulation rate with the ideal rate for non-interacting particles. On the computational side, classical density functional theory (cDFT) simulations were performed to calculate forces between ZnO crystal faces in the presence of a 1 mM zinc acetate dihydrate solution in methanol, encompassing various face configurations and considering electrostatic, ion correlation, and van der Waals forces. Langevin dynamics simulations were also used to model the behavior of two spherical dipoles with an orientation-dependent dipole-dipole interaction potential, aiming to understand the long-range interactions and aligning torques.

The in situ LP-TEM experiments revealed that ZnO nanoparticle growth occurs predominantly through OA. Particle pairs were observed to approach each other, rotate to align their crystallographic axes before reaching contact, and then undergo a jump-to-contact mechanism. The analysis of particle trajectories and orientations showed a clear preference for oriented alignment prior to attachment. The RDF analysis suggested a weak attractive potential extending to separations of several nanometers, indicating an absence of any significant energy barrier to attachment. This finding aligns with the kinetics data, which revealed a negligible barrier to attachment and a diffusion-limited process. The observed diffusion coefficients were significantly lower than predicted by the Stokes-Einstein equation, likely due to interactions with the TEM membrane. The calculated stability ratio supported the conclusion of a diffusion-limited process with negligible attachment barrier. cDFT calculations provided insights into the forces driving OA. While the calculations revealed small, salt-dependent barriers between certain faces, these barriers were insignificant compared to thermal energy (kT). Importantly, long-range attractive interactions and aligning torques were observed extending beyond 5 nm. Langevin dynamics simulations, which incorporated the inherent dipole moment of ZnO, showed that dipole-dipole interactions could contribute significantly to both the long-range attraction and the aligning torques, consistent with experimental observations. The results showed that the timescale for particle rotation is significantly faster than that for translational motion. This allows for particles to align before contact.

The findings address the research question by providing a mechanistic understanding of OA in ZnO nanoparticles. The combination of experimental and computational results reveals that OA is driven by a synergistic interplay of short-range and long-range interactions. Short-range interactions, accurately captured by cDFT simulations incorporating electrostatic, solvation, ion correlation, and van der Waals forces, exhibit minimal barriers to attachment. Long-range interactions, primarily attributed to dipole-dipole forces, are responsible for attracting particles from significant distances and aligning them before they reach the short-range interaction regime. This explains the observed co-alignment prior to contact, a phenomenon not fully understood in previous studies. This study emphasizes the importance of considering both short-range and long-range interactions in understanding OA. While other factors like the electron beam and confinement effects within the TEM fluid cell might have secondary influences, the primary driving forces are clearly linked to the inherent dipole of ZnO and the solution-mediated interactions. This work provides a comprehensive model that successfully explains OA behavior in ZnO, which can serve as a benchmark for understanding OA in other materials, especially polar materials with significant dipole moments.

This study provides novel insights into the energetics and dynamics of oriented attachment. It demonstrates that both short-range and long-range interactions are crucial for OA in ZnO nanoparticles. The findings highlight the significant role of dipole-dipole interactions in driving long-range attraction and pre-alignment of particles before contact. This work contributes to a more comprehensive understanding of OA mechanisms, which can facilitate the design and synthesis of complex materials with tailored morphologies. Future studies should explore other polar materials to test the generality of the findings and further investigate the influence of confinement effects within the TEM fluid cell.

The study primarily focuses on ZnO nanoparticles in methanol solutions under specific conditions. The generalizability of the findings to other materials and solvents needs further investigation. The cDFT model, while effective in capturing several interactions, has limitations at very short distances. Further refinement of the model, such as incorporating charge regulation, might be needed to capture all aspects of the complex interparticle interactions. Also, the quasi-2D nature of the LP-TEM system introduces potential artifacts related to confinement effects, and direct force measurements in bulk solution might provide further validation.