Understanding the fundamental nature of liquids and amorphous solids has long been a "holy grail" of condensed matter physics. Unlike crystals, which possess a periodic lattice, amorphous materials lack long-range order, making their internal structure notoriously difficult to map. For decades, scientists relied on statistical averages from diffraction data, leaving the exact 3D arrangement of atoms to theoretical speculation. This research breaks that barrier, utilizing advanced atomic electron tomography (AET) to provide the first direct experimental determination of 3D atomic positions in monatomic amorphous materials.
The Challenge of Amorphous Structure
In 1952, physicist F.C. Frank hypothesized that atoms in monatomic liquids prefer icosahedral packing—a configuration that is energetically favorable but cannot fill space periodically, leading to "geometric frustration." However, without the ability to see individual atoms in three dimensions, this remained an unproven theory. Traditional methods like X-ray or electron diffraction provide Pair Distribution Functions (PDFs), which describe the average distance between atoms but fail to capture the specific local geometries that define the material's properties.
Atomic Electron Tomography and the RESIRE Algorithm
To overcome these limitations, the researchers employed Atomic Electron Tomography (AET). By capturing multiple 2D projections of a sample at various tilt angles using a Scanning Transmission Electron Microscope (STEM), they could reconstruct a 3D volume.
A significant hurdle in this process was the reconstruction of thin films, where the visible volume changes as the sample tilts. To solve this, the team developed the Real Space Iterative Reconstruction (RESIRE) algorithm. This innovation allowed them to trace and refine individual atomic coordinates with a precision of 18 picometers—roughly one-tenth the size of a typical atom.
Discovery of the Pentagonal Bipyramid Network
The study focused on a Tantalum (Ta) thin film and two Palladium (Pd) nanoparticles. By applying Voronoi tessellation and polytetrahedral analysis to the experimental coordinates, the researchers made a surprising discovery: while icosahedra are rare, the pentagonal bipyramid (a 7-atom motif) is the dominant structural unit.
These bipyramids do not exist in isolation. Instead, they link together by sharing edges to form extensive Pentagonal Bipyramid Networks (PBNs). This represents a previously unrecognized "medium-range order" that bridges the gap between local atomic clusters and the bulk amorphous state. In the Tantalum film, the largest identified PBN consisted of 135 bipyramids, stretching nearly 3 nanometers.
Bridging the Liquid-Glass Divide
To understand how these structures evolve, the team conducted Molecular Dynamics (MD) simulations. These simulations revealed that PBNs are prevalent even in the liquid state at extremely high temperatures. As the liquid is "quenched" (rapidly cooled) into a glass, these networks grow rapidly in size and complexity.
The researchers found a direct correlation between these geometric networks and physical properties: regions with dense PBN packing correspond to higher local mass density. Furthermore, the measured "dihedral angle" between edge-sharing bipyramids was approximately 120.7°, very close to the ideal 120° required to minimize geometric frustration. This explains why these networks can grow significantly without triggering crystallization.
Implications for Material Science
This work provides a new blueprint for understanding the glass transition and the amorphous-crystalline phase transition at the single-atom level. By revealing that the "building blocks" of amorphous solids are interconnected networks rather than isolated icosahedra, the study opens new avenues for engineering metallic glasses with tailored mechanical and thermal properties. The ability to map disordered structures at this resolution marks a transformative step in our ability to design and manipulate materials from the bottom up.
Summary updated