On the connectivity of spaces of three-dimensional tilings

Juliana Freire, Caroline J. Klivans, Pedro H. Milet, Nicolau C. SaldanhaarXiv 2017 · DOI · arXiv

Abstract

We consider domino tilings of three-dimensional cubiculated manifolds with or without boundary, including subsets of Euclidean space and three-dimensional tori. In particular, we are interested in the connected components of the space of tilings of such regions under local moves. Building on the work of the third and fourth authors, we allow two possible local moves, the flip and trit. These moves are considered with respect to two topological invariants, the twist and flux. Our main result proves that, up to refinement: Two tilings are connected by flips and trits if and only if they have the same flux. Two tilings are connected by flips alone if and only if they have the same flux and twist.

Introduction

In two dimensions, the space of domino tilings is relatively well-understood. Most regions can be navigated entirely using a simple local move called a flip—rotating two adjacent parallel dominoes by 90 degrees. However, transitioning to three dimensions introduces significant topological complexity. In a 3D box, tilings can become "stuck," forming fragmented components that cannot reach one another through flips alone. This paper provides a definitive characterization of the connectivity of these 3D tiling spaces by identifying the necessary local moves and the topological invariants that govern them.

A 3D domino tiling of a cubic region and its corresponding 2D projections, illustrating the complexity of mapping 3D configurations.

Methodology: Moves and Invariants

To bridge the gaps between disconnected tiling components, the authors introduce a specifically three-dimensional move called a trit. While a flip involves two dominoes, a trit involves three dominoes within a $2 \times 2 \times 2$ cube. These three dominoes, none of which are initially parallel, are rotated into the only other possible non-parallel configuration.

Two distinct configurations of dominoes within a $2 \times 2 \times 2$ grid, representing the endpoints of a local move.

The connectivity of the space is determined by two primary topological invariants:

  • Flux: Defined using first homology ($H_1$), flux measures the "net flow" of dominoes through the region. It is the fundamental invariant that must be preserved when using both flips and trits.
  • Twist: A more subtle invariant analogous to helicity in fluid dynamics. It measures the topological entanglement or "knottedness" of the dominoes. Twist is preserved by flips but can be altered by trits.

A critical technical component of the proof is refinement. The authors demonstrate that by subdividing each unit cube in the region into a $5 \times 5 \times 5$ grid of smaller cubes, the discrete surfaces required to prove connectivity (Seifert surfaces) are guaranteed to exist. This refinement "unlocks" paths between tilings that might appear isolated in a coarser grid.

Key Results

The central contribution of this work is Theorem 1, which establishes the sufficiency of these invariants for connectivity. For any two tilings $t_0$ and $t_1$ of a region $R$:

  1. Connectivity via Flips and Trits: Tilings are connected (up to refinement) if and only if they share the same Flux.
  2. Connectivity via Flips Alone: Tilings are connected (up to refinement) if and only if they share both the same Flux and the same Twist.

To illustrate this, the authors analyzed a $4 \times 4 \times 2$ box. Despite having approximately 32,000 total tilings, the space is divided into 9 distinct flip-connected components. The largest component (with twist 0) contains over 31,000 tilings, while the remaining components are significantly smaller, sometimes consisting of only a single isolated configuration.

Discussion and Future Work

While the paper successfully characterizes the connectivity of 3D tiling spaces, several questions remain. The necessity of the $5 \times 5 \times 5$ refinement for the proof may not reflect a necessity in practice for simple geometries; proving connectivity on coarser grids remains an open challenge. Additionally, while the invariants are now known, finding the most efficient (shortest) sequence of moves between two tilings is a matter of computational complexity yet to be resolved. Finally, the behavior of tilings in four dimensions and beyond suggests even more complex topological invariants may emerge, marking a clear frontier for future research.

Summary updated

Abstract

We consider domino tilings of three-dimensional cubiculated manifolds with or without boundary, including subsets of Euclidean space and three-dimensional tori. In particular, we are interested in the connected components of the space of tilings of such regions under local moves. Building on the work of the third and fourth authors, we allow two possible local moves, the flip and trit. These moves are considered with respect to two topological invariants, the twist and flux. Our main result proves that, up to refinement: Two tilings are connected by flips and trits if and only if they have the same flux. Two tilings are connected by flips alone if and only if they have the same flux and twist.
BibTeX
@misc{freire2017connectivity,
  author = {Juliana Freire and Caroline J. Klivans and Pedro H. Milet and Nicolau C. Saldanha},
  title = {{On the connectivity of spaces of three-dimensional tilings}},
  year = {2017},
  eprint = {1702.00798},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1702.00798},
}

Loading related content...

Figure