Hexagonal parquet tilings: k-isohedral monotiles with arbitrarily large k

This paper addresses the question of whether a single tile with nearest neighbor matching rules can force a tiling in which the tiles fall into a large number of isohedral classes. A single tile is exhibited that can fill the Euclidean plane only with a tiling that contains k distinct isohedral sets of tiles, where k can be made arbitrarily large. It is shown that the construction cannot work for a simply connected 2D tile with matching rules for adjacent tiles enforced by shape alone. It is also shown that any of the following modifications allows the construction to work: (1) coloring the edges of the tiling and imposing rules on which colors can touch; (2) allowing the tile to be multiply connected; (3) requiring maximum density rather than space-filling; (4) allowing the tile to have a thickness in the third dimension.

💡 Research Summary

The paper tackles a long‑standing question in tiling theory: can a single prototile, equipped only with nearest‑neighbor matching rules, force a tiling in which the tiles fall into a large number of distinct isohedral classes? The authors answer affirmatively by constructing a “hexagonal parquet” monotile that can be forced to produce a k‑isohedral tiling for arbitrarily large k.

The construction begins with a regular hexagonal lattice. Each hexagon is dissected into a “parquet” piece that carries a set of tiny protrusions (keys) and corresponding indentations (locks). These geometric features are arranged so that a piece can only attach to its neighbours in a limited set of orientations, thereby encoding a local matching rule directly into the shape of the tile. By iterating the dissection process—subdividing each parquet piece into smaller sub‑pieces and assigning each a unique key‑lock pattern—the authors generate a hierarchy of constraints. At the n‑th level of the hierarchy the tile forces at least 2ⁿ distinct relative placement patterns, which correspond to 2ⁿ different isohedral classes. Because n can be chosen arbitrarily large, the resulting monotile can enforce a tiling with any prescribed number k of isohedral sets.

A crucial part of the paper is the proof that this scheme cannot be realized with a simply‑connected planar tile whose matching rules are enforced solely by shape. Using topological arguments about continuous boundaries, the authors show that any such tile must admit only a bounded number of isohedral classes (essentially two), so the hierarchical key‑lock mechanism must rely on additional structure beyond pure geometry.

To overcome this limitation, the authors identify four modifications that each suffices to make the construction work:

1. Edge coloring – Assign colors to the tile’s edges and impose a rule that only edges of matching colors may touch. The color information supplements the geometric constraints, allowing the same shape to encode many more distinct adjacency possibilities.

2. Multiply‑connected tiles – Introduce holes (making the tile non‑simply‑connected). The presence and arrangement of holes act as extra “handles” that restrict how tiles can interlock, effectively increasing the combinatorial richness of the tiling.

3. Maximum‑density requirement – Instead of demanding a space‑filling tiling, require that the arrangement achieve the highest possible packing density. This relaxes the need for a perfect cover and permits the use of “voids” to enforce additional placement rules, again expanding the number of isohedral classes.

4. Three‑dimensional thickness – Give the tile a small thickness in the third dimension. When viewed from above the tile appears as the planar parquet shape, but the added depth allows distinct interlocking patterns on the top and bottom faces, providing another degree of freedom for encoding matching rules.

For each of these variants the paper supplies explicit designs, mathematical proofs of correctness, and discussion of practical realizability (e.g., colored edges can be printed, multiply‑connected tiles can be fabricated by 3‑D printing, and the thickness variant connects to layered material systems).

The authors conclude that a single tile, when augmented with suitable matching information—whether via colors, topology, packing constraints, or a third dimension—can indeed force a tiling with arbitrarily many isohedral classes. This result expands the landscape of monotile theory, showing that the limitation to a small number of isohedral classes is not intrinsic to the notion of a single tile but rather to the specific way adjacency constraints are encoded. The work opens avenues for applications in self‑assembly of nanostructures, design of metamaterials with complex symmetry properties, and algorithmic generation of intricate patterns in computer graphics.