Ice Defects Explain Common Honeycomb Patterns Seen under Microscopes

Scientists have long sought to understand the subtle structural defects within ice that influence its properties. Jingshan S. Du, Suvo Banik from the Center for Nanoscale Materials at Argonne National Laboratory, Lehan Yao and Shuai Zhang from the Physical Sciences Division at Pacific Northwest National Laboratory, working with colleagues including Subramanian K. R. S. Sankaranarayanan from Argonne and James J. De Yoreo from Pacific Northwest National Laboratory, now demonstrate that the commonly observed honeycomb pattern in hexagonal ice (Ih) originates not from individual oxygen columns, but from intrinsic basal stacking faults. Their research, utilising phase-contrast transmission electron microscopy at a record-breaking line resolution of 89 picometers, finer than the length of an oxygen-hydrogen covalent bond, reveals a translational boundary between domains causing the observed contrast. This discovery clarifies the structural relationships between hexagonal, stacking-disordered and cubic ice phases, and represents a significant advance in our ability to characterise the complex molecular packing of water in its solid state.

Understanding these fundamental structures is crucial for fields ranging from cryobiology to materials science. This feat, accomplished using a novel cryogenic liquid-cell transmission electron microscopy (CRYOLIC-TEM) technique, has revealed that the commonly observed honeycomb patterns in hexagonal ice (Ih) are not indicative of neatly arranged oxygen columns as previously thought.

Instead, these patterns originate from intrinsic basal stacking faults, subtle imperfections in the way water molecules pack together within the ice structure. Translational boundaries, where domains of ice are slightly offset from one another by (2/3a + 1/3b) in-plane, create this honeycomb-like contrast in phase-contrast transmission electron microscopy (TEM) images.

This discovery challenges conventional interpretations of HRTEM images of ice and provides a more accurate understanding of its molecular arrangement. Domains with nonequivalent translations produce patterns resembling cubic ice (Ic) but with a 3-fold symmetry, clarifying the structural relationships between different ice phases. The ability to resolve these stacking faults highlights the remarkable defect tolerance of ice’s molecular packing, a property crucial for understanding its behaviour in diverse environments.

The team’s work not only refines our knowledge of ice’s fundamental structure but also establishes a new benchmark for high-resolution imaging of delicate materials. The CRYOLIC-TEM method protects crystalline ice thin films from electron beam damage and vacuum conditions, enabling prolonged observation at the atomic scale. By encasing the ice between amorphous carbon membranes, researchers maintained high-quality samples for extended periods, allowing for detailed analysis of structural perturbations.

This breakthrough in sample preparation, combined with aberration-corrected HRTEM, facilitated the observation of these subtle stacking faults and the subsequent determination of their origin. The findings have significant implications for fields ranging from materials science to glaciology, offering new avenues for characterising the behaviour of water in the solid state.

This resolution milestone opens possibilities for investigating subtle structural changes in ice and other materials with unprecedented precision. The ability to visualize defects at this scale will undoubtedly contribute to a deeper understanding of crystallization mechanisms and intermolecular interactions, potentially leading to advancements in areas such as cryopreservation and the design of novel materials. The research underscores the power of advanced microscopy techniques to reveal the hidden complexities of even the most familiar substances.

Cryogenic liquid-cell TEM enables atomic-resolution imaging of stable ice crystals

Researchers employed aberration-corrected high-resolution transmission electron microscopy (HRTEM), specifically utilising a FEI Titan HRTEM corrected for image-forming lenses coupled with a Gatan Metro 300 direct electron detector operating in electron-counting mode, to achieve this level of detail. This instrumentation facilitated the acquisition of high-quality phase-contrast data essential for discerning subtle structural features within ice.

A novel sample preparation method, termed cryogenic liquid-cell TEM (CRYOLIC-TEM), was central to the study’s success. CRYOLIC-TEM protects crystalline ice thin films from the damaging effects of high vacuum and electron beam exposure by encapsulating samples between amorphous carbon membranes. This technique allows for extended observation, crystals remained stable under HRTEM imaging for minutes, and enables in situ observation of dynamic processes within the ice lattice.

Researchers initially prepared ice thin films using both vitrification and direct deposition from residue vapor within the TEM column onto lacey carbon or graphene supports.

To investigate the observed phase-contrast patterns, researchers systematically imaged hexagonal ice (Ih) along the c-axis, focusing on two frequently observed arrangements: a hexagonal dot array and a honeycomb structure. Time-series images, spatially registered and recorded under constant electron flux, were captured to track the evolution of these patterns and determine the mechanisms driving transitions between them.

Kinematic TEM image simulations were performed to model the expected contrast variations under different conditions, providing a theoretical framework for interpreting the experimental observations. Atomic force microscopy (AFM) was also used to characterise the supporting carbon membranes, quantifying height variations to rule out artefacts arising from sample topography.

Domain boundaries define honeycomb patterns in hexagonal ice

Achieving a line resolution of 89 picometers, finer than the oxygen-hydrogen covalent bond length, the research detailed the structural characteristics of hexagonal ice. This unprecedented resolution enabled the observation of intrinsic basal stacking, revealing that honeycomb-like patterns previously attributed to individual oxygen columns actually originate from translational boundaries between domains.

These domains exhibit an in-plane offset of 2/3a + 1/3b, creating the observed contrast in phase-contrast transmission electron microscopy. Simulations of ice crystals with varying ratios of AB- and BC-domains demonstrated that the presence of either stacking fault generates a 3-fold-symmetric honeycomb pattern. The intensity of the dots within these honeycombs is tunable, dependent on both the thickness ratio of the domains and the defocus value employed during imaging.

This tunability aligns with the transitional behaviour observed experimentally, suggesting that the formation of these patterns requires only an intrinsic basal stacking fault. The energy required for such faults is less than 1 meV/atom, a value significantly below thermal fluctuations at both 0°C (approximately 23.5 meV) and −180°C (approximately 8 meV), indicating their ease of formation.

Phase-contrast CRYOLIC-TEM imaging revealed a 3-fold-symmetric pattern arising from vertically stacked AB, BC, and AC domains along the c axis. Spatial averaging identified three groups of bright dots with differing intensities within repeating unit cells, matching the structural model of these stacked domains. Modelling a 3-layer crystal of translational domains, BC|AB|AC with thicknesses of 11, 8, and 6 unit cells, replicated this 3-fold symmetry pattern in silico, further validating the experimental observations. This arrangement corresponds to “stacking disordered” type-I ice, a structure often found alongside imperfect cubic type-I ice exhibiting mixtures of Ih and Ic domains.

Revealing basal stacking faults clarifies hexagonal ice structure and behaviour

The latest work, achieving sub-angstrom resolution in ice imaging, represents a significant step towards the long-held goal of scientists to truly ‘see’ water, not just as a fluid but in its solid forms at the molecular level. For decades, interpreting transmission electron microscopy images of ice has been hampered by ambiguity. This research decisively demonstrates that the commonly observed honeycomb patterns in hexagonal ice aren’t evidence of perfect order, but rather the result of basal stacking faults, essentially, slight misalignments in the layers of molecules.

The implications extend beyond simply correcting a misinterpretation. Understanding these defects is crucial because they influence the behaviour of ice, impacting everything from glacier formation to the stability of biological molecules preserved in frozen states. The ability to visualise these features with such precision opens doors to characterising other subtle structural perturbations in ice and related materials, potentially revealing how water interacts with proteins or other compounds within icy environments.

Scaling this method to study more complex, disordered ice systems or real-world samples remains a considerable challenge. Future work will likely focus on automating the image analysis process and extending the technique to in-situ studies, observing how ice structure evolves under varying temperature and pressure conditions. Ultimately, this achievement isn’t just about seeing smaller details; it’s about building a more complete and accurate picture of one of the most fundamental substances on Earth.