Voronoi Geometry for Local Structural Analysis

This direction is merged into another direction: Voronoi-based spatial partitioning. Its last version stays readable and citable here.

Version 1 of 116 papers21 conceptsfrozen 2026-10-04

In the study of disordered matter—from the chaotic packing of sand grains to the "liquid-like" arrangement of atoms in glass—the central challenge is defining what a "neighbor" actually is. Traditional methods often rely on a fixed distance cutoff: if two atoms are within 3 Ångströms, they are connected. However, this approach is arbitrary and fails to capture the physical reality of volume, surface contact, and multi-body interactions.

Researchers across materials science, biology, and machine learning have converged on a more robust geometric solution: Voronoi tessellation. By partitioning space into polyhedral cells based on proximity, this framework provides a parameter-free way to quantify the local environment of any particle.

The Intuition of the Voronoi Cell

Imagine a crowded room where every person claims all the floor space that is closer to them than to anyone else. The boundaries between people would form a network of lines; in three dimensions, these boundaries become the faces of a polyhedron. This is a Voronoi cell.

For a system of particles, the Voronoi cell of a specific atom represents its "territory." This geometric construction offers several immediate advantages over simple distance metrics:

  1. Volume as a Fundamental Variable: The volume of the cell represents the "free volume" or "rattle room" available to a particle. Chaoming Song, Yuliang Jin, and Hernán A. Makse demonstrated that in jammed matter, this volume acts as a conserved quantity analogous to energy, bridging the gap between microscopic arrangements and macroscopic packing limits in Jamming I: A volume function for jammed matter.
  2. Physical Contacts: A shared face between two Voronoi cells indicates a physical neighbor. The area of that face quantifies the strength or extent of the interaction, a concept Fei Guo and colleagues used to identify protein-protein binding sites more accurately than distance-based methods in Structural neighboring property for identifying protein-protein binding sites.
  3. Shadowing and Solid Angles: Voronoi cells capture how neighbors "shield" an atom from others. By calculating the solid angle subtended by a Voronoi face, researchers like Mohammadreza Karamad can represent the 3D "shadowing" effects of local geometry, which is critical for predicting material properties in Orbital Graph Convolutional Neural Network for Material Property Prediction.

Finding Order in Amorphous Solids

One of the most significant breakthroughs using this approach occurred in the study of amorphous (non-crystalline) solids. For decades, these materials were thought to be almost entirely disordered. However, by applying Voronoi tessellation to 3D atomic coordinates obtained through atomic electron tomography, Yakun Yuan, Jihan Zhou, and Dennis Kim revealed a hidden "medium-range order."

In Direct observation of 3D atomic packing in monatomic amorphous materials, they showed that amorphous tantalum and palladium are not random; they are built from pentagonal bipyramid networks (PBNs). By analyzing the symmetry and connectivity of Voronoi polyhedra, they could classify these local motifs and show how they link together to form the material's backbone. This structural "smoking gun" explains why certain liquids transition into glasses rather than crystals: the local Voronoi geometry (like five-fold symmetry) is physically incompatible with long-range crystalline tiling.

From Granular Flow to Protein Folding

The utility of Voronoi analysis extends from the sub-atomic to the macroscopic. In granular materials, where particles are often of different sizes (polydisperse), a standard Voronoi diagram fails because it doesn't account for particle radii. Zhifeng Li and Jie Zhang utilized Radical Voronoi tessellation—a weighted version of the diagram—to study how different-sized grains segregate under shear in Granular Segregation Mechanisms by Cyclic Shear. This allowed them to track local density asymmetries that drive large particles to the surface.

In the biological realm, Ilia Igashov and Česlovas Venclovas developed VoroCNN, a deep learning framework that replaces traditional 3D voxels with Voronoi cells. By representing a protein as a graph of Voronoi polyhedra, the model captures physically meaningful attributes like atomic contact areas and multi-body interactions that are lost in grid-based discretizations.

Why It Works: The Geometric Essence

The success of Voronoi-based methods stems from their ability to provide a canonical description of space. As Marco M. Mosca argues in Voronoi-based similarity distances between arbitrary crystal lattices, using Voronoi cells avoids the ambiguities of traditional unit cells, which can be defined in many different ways for the same lattice. The Voronoi cell is unique to the arrangement of points.

Whether it is mapping the free-energy landscape of $CO_2$ in water (Daniela Polino et al.) or ensuring the interpretability of deep learning models for materials (Tien-Sinh Vu), Voronoi tessellation serves as a mathematical bridge. It transforms a set of raw coordinates into a structured, physically relevant map of local influence, revealing the hidden geometric logic of matter.

Papers in this version

History & citationVersion 1 · 2026-10-04

Merged into another direction · Permanent link

View frozen version 1

Cite this direction

Latest version (a living page):

@misc{lacuna_voronoi_geometry_for_local_structural_an_cd7f89b1,
  title = {Voronoi Geometry for Local Structural Analysis},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/voronoi-geometry-for-local-structural-analysis/rdr_cd7f89b1fc0549da993e9a18118edbf6},
  note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-19},
}

This exact version (version 1, frozen text):

@misc{lacuna_voronoi_geometry_for_local_structural_an_cd7f89b1_v1,
  title = {Voronoi Geometry for Local Structural Analysis},
  author = {{Lacuna}},
  howpublished = {Lacuna research direction},
  year = {2026},
  url = {https://lacuna.tiptreesystems.com/direction/voronoi-geometry-for-local-structural-analysis/rdr_cd7f89b1fc0549da993e9a18118edbf6/v/1},
  note = {Version 1, frozen 2026-10-04. Accessed 2026-11-19},
}