CryoTEM as a Tool for Investigating Crystallization in Aqueous Systems

Solution based synthesis is a common method for production of energy-related materials and a pervasive process in environmental settings tied to energy technology through the impact on carbon emissions and climate. The realization that crystallization in aqueous systems occurs along diverse structural pathways has motivated extensive TEM-based investigations to understand the nature of those pathways, the associated dynamics, and the underlying thermodynamic and kinetic factors. Ex situ TEM studies of such systems suffer from the unknown impact of drying, as well as an inability to directly relate structures seen at one time point with those present at another, rendering an understanding of pathways nearly impossible. For systems amenable to in situ liquid phase TEM, these two problems can be readily circumvented, but the effect of the electron beam through the creation of radiolysis products that alter the solution chemistry and, in the case of organic compounds, the breakdown of the solutes themselves often cast doubt on the applicability of the results to synthetic processes outside the microscope. CryoTEM offers a middle ground that gives a true picture of crystallization at any given time point and allows for the construction of likely pathways per the ergodic principle by analyzing structures present over large areas as a substitute for following individual structures over time. Here we present two examples of crystallization processes for which cryoTEM provides critical information concerning the structural state. The first involves the process of crystal growth through repeated oriented attachment (OA) events, i.e., the growth of single crystals by repeated fusion of nanocrystals that adopt crystallographic coalignment prior to attachment. Using data from minerals aluminum and iron oxide growing by OA, we show that cryogenic high-resolution TEM (HRTEM) can provide definitive information on the temporal evolution of the growing crystals and the structural relationship of the primary particles. The second example involves the crystallization of the water itself. Visualizing the crystalline network of ice with atomic resolution in real space is challenging due to the weak directional hydrogen bonds between water molecules and the resulting instability of ice under an electron beam in vacuo. However, we find that crystallizing water encapsulated between amorphous carbon membranes and freezing it in liquid nitrogen to form hexagonal ice (Ih ice) enables successful HRTEM of the resulting ice sections. Even though the sections show overall single-crystallinity by diffraction standards, the crystal may exhibit structural variation and contain sub-domains ~ 10 nm in size. Quantitative strain and tilt mapping reveals the sub-surface flexibility of ice and is used to demonstrate the correlation between defect structures and strain accumulation. Both sharp low-angle grain boundaries and gradual lattice bending are associated with sub-domains with different crystal orientations. Taken together, this work shows promise in elucidating the molecular and defect structures of ice and may shed light on the physics of ice surfaces and their interfaces with substrates that promote ice nucleation.