🔬 condensed matter

Prismatic Soft Cubes

This paper extends existing methods for finding soft tilings to the cubic lattice by imposing natural geometric constraints rather than symmetry requirements, resulting in the classification of 26 distinct soft cubic cells and 68 fundamental domains organized into 8 symmetry groups, supported by a new Python-based algorithmic process.

Original authors: Kinga Kocsis

Published 2026-08-04

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Kinga Kocsis

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are building a giant, infinite castle out of blocks. In the world of geometry, these blocks are called "tiles," and when they fit together perfectly to cover a floor or fill a room without any gaps or overlaps, it's called a "tiling." For centuries, mathematicians have loved the perfect cube: a block with six flat square faces and eight sharp, pointy corners. It's the ultimate Lego brick. But what if you wanted to build a castle where the blocks didn't have any sharp, painful corners? What if every edge was smooth, like a river stone, yet the blocks still fit together perfectly to fill the entire universe?

This is the puzzle of "soft cells." In geometry, a "sharp corner" is a point where no smooth curve can pass through; it's a jagged edge. A "soft" shape is the opposite: at every single point on its surface, you can draw a smooth, curvy line. The big question is: Can you take a standard grid of cubes and "soften" them—rounding off their edges and corners—so they still fit together like a jigsaw puzzle, but without any of the sharp points? It turns out this isn't just a fun math game; it helps us understand how nature packs things together, from the way crystals grow to how cells in our bodies might arrange themselves.

In this paper, a mathematician named K. Kocsis takes on the challenge of softening the most basic grid of all: the cubic lattice. Think of the cubic lattice as a 3D grid of invisible points, like the corners of a giant stack of dice. The author asks: If we keep the grid points in the same place and only smooth out the edges of the cubes, how many different ways can we do this? The paper proves that there are exactly 26 unique ways to create these "soft cubes" that fit together perfectly.

Here is the twist: The author didn't just guess these shapes. They used a strict set of rules to find them. First, the "half-tangents" (the directions the edges point as they leave a corner) had to stick to the original grid lines, like train tracks. Second, the edges had to be "planar," meaning they lay flat on a surface rather than twisting in 3D space like a corkscrew. Under these specific conditions, the paper proves that only 26 distinct soft cube shapes exist.

The paper also sorts these 26 shapes into two camps. Some of them are "achiral," meaning they are their own mirror images; you can flip them over, and they look the same. Others are "chiral," like your left and right hands. If you have a chiral soft cube, you can't just slide it around to fill the space; you need its mirror image (the "left hand" and the "right hand") to work together to fill the gaps. Out of the 26 shapes, 9 are achiral, and 17 are chiral.

But the story doesn't stop at the single block. The author also figured out how these blocks stack up to form the smallest repeating unit, called a "fundamental domain." By mixing and matching these 26 soft cubes, the paper shows you can build 68 different fundamental domains. These domains can be grouped into 8 different families based on their symmetry, which relates to how they fit into the crystal systems of triclinic, monoclinic, orthorhombic, and tetragonal.

One of the most important things the paper does is rule out a common hope. You might think that if you soften a cube, it might still keep all the perfect symmetry of the original cube. The paper proves this is impossible. None of the 26 soft tilings share the full symmetry group of the original cubic lattice. The act of smoothing the corners inevitably breaks some of the perfect symmetry.

To find these numbers, the author didn't just draw pictures; they wrote a computer program (in Python) to act as a digital detective. The program checked thousands of combinations of curved faces and edges, filtering out the duplicates and the impossible shapes. It confirmed that while there are many ways to bend the edges, only 26 specific configurations result in a shape that is both soft and space-filling.

So, the next time you see a stack of boxes, imagine if they were made of soft, squishy clay that still held their shape perfectly. This paper tells us that there are exactly 26 ways to sculpt that clay into a cube that fits into a perfect grid, and it gives us the map to find them all. It's a precise, mathematical exploration of how we can make the rigid world a little bit softer, without losing the perfect fit.

Technical Summary: Prismatic Soft Cubes

Problem Statement
The paper addresses the geometric problem of "softening" a standard cubic lattice tiling. A soft cell is defined as a space-filling shape without sharp corners, meaning that at every surface point, at least one smooth curve passes through. While previous research established that algorithms exist to soften convex polyhedral tilings while preserving their combinatorial structure and lattice points, these proofs did not characterize all possible soft tilings for a given lattice. Specifically, for the cubic lattice (where unit cubes meet at nodes), the paper seeks to determine the complete set of soft tilings that agree to the first order with the cubic lattice (preserving node locations and face structure) under specific geometric constraints: the edge half-tangents must be restricted to lattice directions, and the edges must be planar ( $C^1$ planar curves).

Methodology
The author employs a multi-stage analytical and algorithmic approach to classify these soft tilings:

Node Softening Analysis: The process begins by analyzing the "nodal set" of six edge half-tangent vectors meeting at a cubic lattice node. Using the Extended Edge Bending (EEB) algorithm, the author derives "softening equations" where pairs of half-tangents must satisfy $\mathbf{u}_i \cdot \mathbf{u}_j = -1$ to ensure smoothness at the corners.
- Theorem 1 establishes that there are 12 complete sets of softening equations for parallel soft tilings.
- Theorem 2 narrows this down, proving that only one specific configuration (and its 3 equivalent rotations) ensures the softness of all eight corners meeting at a node. This configuration requires the half-tangents to form three coincident pairs of opposite vectors.
Prismatic Structure Proof: Theorem 3 proves that any soft tiling satisfying these conditions must possess a prismatic structure. This implies the existence of a projection direction where all perpendicular cross-sections form a square lattice, and the cells have four parallel, congruent straight edges.
Corner and Face Classification: With the nodal configuration fixed, the paper classifies the possible curved edges and faces.
- Edges: Four distinct curved edge types are identified based on the relative directions of half-tangents at their endpoints (identical or opposite, and their mirror images).
- Faces: These edge types combine to form six distinct curved face types (labeled I through VI).
- Cells: The author constructs soft cubic cells by pairing these curved faces. A manual algorithm (Algorithm 1) and a Python-based algorithm (Algorithm 2) are used to enumerate all possible pairings, eliminate duplicates via rotation and reflection, and verify space-filling properties.
Fundamental Domain Generation: For each valid soft cell, the paper constructs "fundamental domains" (minimal sets of cells that can be translated to fill space without gaps or overlaps). This involves analyzing translational compatibility in the $x$ , $y$ , and $z$ directions, considering reflections and rotations required to close the tiling.

Key Contributions and Results

Existence of 26 Soft Cubes: The primary result (Theorem 4) is the proof that there are exactly 26 distinct parallel soft tilings that partially agree to the second order with the cubic lattice. These are classified by the types of curved faces they possess (e.g., I-I, I-III, IV-VI).
Symmetry and Chirality: The study distinguishes between achiral tilings (fillable by translation and rotation alone) and chiral tilings (requiring mirror images). Of the 26 cells, 9 yield achiral tilings, while 17 are chiral.
Fundamental Domains: The paper demonstrates that these 26 cells generate a total of 68 distinct fundamental domains. These domains are classified into 8 groups based on their lattice symmetry (triclinic, monoclinic, orthorhombic, and tetragonal systems).
Symmetry Limitations: A significant finding is that none of the 26 soft tilings preserve the full symmetry group of the original cubic lattice. The richest symmetry groups found among the solutions are of order eight (e.g., $D_{2d}$ , $D_{2h}$ , $C_{4h}$ ).
Algorithmic Implementation: The paper provides a detailed description of the classification algorithms and notes the availability of a Python implementation to reproduce the classification of the 26 non-equivalent geometric cell types.

Significance
The paper claims to extend previous methods used for truncated octahedral tilings to the cubic lattice, removing the necessity for imposed symmetry constraints. By imposing only "natural conditions" (lattice-aligned half-tangents and planar edges), the author provides a complete enumeration of the geometric possibilities for soft cubic cells. The work clarifies the geometric constraints required for space-filling soft cells and establishes that while softening is possible, it inherently breaks the full symmetry of the cubic lattice. The results offer a comprehensive catalog of prismatic soft cells and their associated fundamental domains, contributing to the broader understanding of space-filling geometries in convex geometry and crystallography.

More like this