The volume of hyperbolic alternating link complements

Marc Lackenby

Abstract

If a hyperbolic link has a prime alternating diagram $D$, then we show that the link complement's volume can be estimated directly from $D$. We define a very elementary invariant of the diagram $D$, its twist number $t(D)$, and show that the volume lies between $v_3(t(D) - 2)/2$ and $v_3(16t(D) - 16)$, where $v_3$ is the volume of a regular hyperbolic ideal 3-simplex. As a consequence, the set of all hyperbolic alternating and augmented alternating link complements is a closed subset of the space of all complete finite volume hyperbolic 3-manifolds, in the geometric topology. The appendix by Ian Agol and Dylan Thurston improves the upper bound on volume to $v_3(10t(D) - 10)$.

Introduction

In knot theory, a central challenge is relating a knot's basic topological description—how it is drawn in 2D—to its geometric structure in 3D space. For many knots, the space around them (the link complement) can be assigned a unique hyperbolic geometry. The most fundamental invariant of this geometry is its volume. Before this work, calculating the volume of a hyperbolic link complement usually required intensive software or complex manifold decompositions. This paper bridges that gap by proving that the hyperbolic volume of an alternating link complement is linearly bounded by its "twist number," providing a direct way to estimate a complex geometric invariant from a simple diagram.

Defining the Twist Number

The core of this estimation method is the twist number $t(D)$ of a diagram $D$. A "twist" is defined as a maximal sequence of crossings between two strands (forming bigons) or a single isolated crossing. This invariant captures the geometric "bulk" of the link more effectively than a simple crossing count. To ensure the metric is robust, diagrams are often assumed to be "twist-reduced," preventing artificial inflation of the count through unnecessary complexity.

Figure 1: A link diagram illustrating the identification of twists. Here, the twist number $t(D) = 3$.

Methodology: Bounding the Volume

Lackenby employs two distinct strategies to establish the volume bounds:

Upper Bound via Gromov Norm

The upper bound is derived using techniques related to the Gromov norm. By performing "Dehn filling" (a surgery-like operation) on a simpler "augmented" link, the volume of the original knot is shown to be strictly less than the volume of a link complement that can be decomposed into a specific number of ideal tetrahedra. Since the volume of any ideal tetrahedron is capped at $v_3 \approx 1.01494$, the total volume is bounded by the number of tetrahedra.

Lower Bound via Checkerboard Surfaces

The lower bound utilizes a theorem regarding the "guts" of incompressible surfaces. By analyzing the "checkerboard surfaces"—the black and white regions of an alternating diagram—the Euler characteristic is used to establish a minimum "floor" for the volume. This reveals that the topology of the regions between strands directly determines the geometric bulk of the manifold.

Figure 4: Visualization of checkerboard surfaces and their orientation, used to determine the "guts" of the manifold and establish a lower volume bound.

Key Results

The primary result is Theorem 1, which states that for a prime alternating diagram $D$ of a hyperbolic link $K$, the volume is bounded by: $$\frac{v_3(t(D) - 2)}{2} \leq \text{Volume}(S^3 - K) < v_3(16t(D) - 16)$$

An appendix by Ian Agol and Dylan Thurston tightened the upper bound significantly to: $$\text{Volume}(S^3 - K) < 10v_3(t(D) - 1)$$

The authors further demonstrate that this $10v_3$ bound is "asymptotically sharp," meaning there are sequences of knots (such as chain link fence links) where the volume-per-twist ratio approaches this limit as the number of twists goes to infinity.

Conclusion and Future Work

This research reveals that groupings of crossings (twists) are the fundamental units dictating geometric volume in alternating links. While the upper bound applies broadly, the tight lower bound relies heavily on alternating properties. Future work may aim to extend these bounds to non-alternating hyperbolic knots and further narrow the gap between the lower and upper constants.

Summary updated

Abstract

If a hyperbolic link has a prime alternating diagram D, then we show that the link complement's volume can be estimated directly from D. We define a very elementary invariant of the diagram D, its twist number t(D), and show that the volume lies between v_3(t(D) - 2)/2 and v_3(16t(D) - 16), where v_3 is the volume of a regular hyperbolic ideal 3-simplex. As a consequence, the set of all hyperbolic alternating and augmented alternating link complements is a closed subset of the space of all complete finite volume hyperbolic 3-manifolds, in the geometric topology. The appendix by Ian Agol and Dylan Thurston, which was written after the first version of this paper was distributed, improves the upper bound on volume to v_3(10t(D) - 10). In addition, examples of alternating links are given which asymptotically achieve this bound.
BibTeX
@inproceedings{lackenby2004volume,
  author = {Marc Lackenby},
  title = {{The volume of hyperbolic alternating link complements}},
  booktitle = {Proceedings of the London Mathematical Society},
  year = {2004},
  doi = {10.1112/s0024611503014291},
}

Loading related content...

Figure