Abstract
An infinite class of $\mathcal{N}=1$ gauge theories can be engineered on the worldvolume of D3-branes probing toric Calabi–Yau 3-folds. This kind of setup has multiple applications, ranging from the gauge/gravity correspondence to local model building in string phenomenology. Brane tilings fully encode the gauge theories on the D3-branes and have substantially simplified their connection to the probed geometries. The purpose of this paper is to push the boundaries of computation and to produce as comprehensive a database of brane tilings as possible. We develop efficient implementations of brane tiling tools particularly suited for this search. We present the first complete classification of toric Calabi–Yau 3-folds with toric diagrams up to area 8 and the corresponding brane tilings. This classification is of interest to both physicists and mathematicians alike.
Introduction
In string theory, engineering 4D particle physics theories often involves placing D3-branes at the singularity of a six-dimensional toric Calabi-Yau (CY) 3-fold. The resulting gauge theory is elegantly described by a brane tiling (or dimer model)—a bipartite graph on a torus. These tilings act as a "Rosetta Stone," mapping geometric data to physical parameters: faces correspond to gauge groups, edges to chiral fields, and nodes to superpotential terms.
While the theoretical link is established, calculating tilings for complex geometries is computationally intensive. Previous approaches like the "Inverse Algorithm" faced exponential scaling issues. This work addresses this gap by providing a systematic, large-scale database of these models, crucial for identifying string configurations that align with observed physical phenomena.
Methodology: Partial Resolution and Automation
The authors utilize Partial Resolution to bypass the limitations of previous algorithms. This process begins with a highly symmetric "parent" theory, such as an orbifold of the conifold, which possesses a simple square-lattice brane tiling. By assigning vacuum expectation values (vevs) to specific fields—a process known as "higgsing"—edges in the tiling are removed. Geometrically, this corresponds to deleting points from the toric diagram.
To scale this approach, the authors developed a suite of Mathematica modules. These tools automate the identification of field combinations required to reach target geometries while ensuring physical consistency. Key graph-theoretic tests verify that the number of gauge groups in the resulting tiling matches the area of the target toric diagram.
Results and Dataset
The primary output is a comprehensive classification of all $SL(2, \mathbb{Z})$ inequivalent convex lattice polygons (toric diagrams) up to area 8. The area of a toric diagram is directly proportional to the number of gauge groups in the 4D theory, serving as a proxy for model complexity.
The database includes:
- Area 6: 13 independent toric diagrams.
- Area 7: 11 independent toric diagrams.
- Area 8: 27 independent toric diagrams.
For each of these 51 geometries, the paper provides the toric diagram, bipartite brane tiling, quiver diagram, and explicit superpotential. The accompanying Mathematica toolkit (RecDimerModels, ToricInfo, etc.) allows researchers to extend these results to even larger areas.
Discussion and Future Directions
A key insight from this study is the requirement for maximal higgsing to maintain tiling consistency; insufficient edge removal can lead to theories that do not accurately represent the target geometry. While the current database provides at least one consistent tiling per geometry, many geometries possess multiple "toric phases" related by Seiberg duality. Future work aims to classify these phases and incorporate additional physical data like R-charges. Ultimately, this library serves as a foundation for "local model building," where researchers seek to embed Standard Model gauge groups into these toric frameworks to find realistic string vacua.
Summary updated