Tight contact structures on the Whitehead link surgery family $M(n,r)$: classification beyond the resolved parameter range and universal tightness

Min and Nonino (arXiv:2311.14103, published in Geometriae Dedicata 220, 2026) classified, up to isotopy, the tight contact structures on $M(n,r)$, the result of $(n,r)$-surgery on the Whitehead link — the first such classification for an infinite family of hyperbolic L-spaces — for integers $n\ge5$ and slopes $r\in\mathcal{R}_+\cup\mathbb{Q}_-$, proving every such structure is Stein fillable and most are virtually overtwisted. Two coherent questions remain open. First, outside this parameter range (integers $2<n<5$, or $r\notin\mathcal{R}_+\cup\mathbb{Q}_-$; excluding the slope $r=0$, where $M(n,0)$ is a torus bundle over $S^1$ already known to carry infinitely many tight contact structures) only the lower bounds of $\Psi(r)+2\Phi(r)$ (for $r\ge1$) or $\Psi(r)$ (for $r<1$) structures are known, and the authors state there "could be more": no upper bound or classification exists. Second, in the small-count regimes of their Remark 1.8 (e.g. integer slopes $r\ge2$ with $\Psi(r)+2\Phi(r)<6$) it is unknown whether any tight structure on $M(n,r)$ is universally tight or whether all are virtually overtwisted. The general conjecture that motivated this (their Conjecture 1.1: a hyperbolic 3-manifold is an L-space if and only if it supports no universally tight contact structure) was refuted in 2025 by Baldwin, Sivek, and Zung, who constructed hyperbolic L-spaces carrying universally tight contact structures; since their examples do not lie in the Whitehead link family, the universal-tightness question for the specific manifolds $M(n,r)$ survives as a concrete open problem decoupled from the refuted conjecture.

Background

A contact structure $\xi$ on an oriented 3-manifold $M$ is a maximally nonintegrable plane field; it is overtwisted if $M$ contains an overtwisted disk (an embedded disk tangent to $\xi$ along its boundary) and tight otherwise. Eliashberg's classification reduced overtwisted structures to homotopy data, so the hard problem is classifying tight ones. A tight $\xi$ is universally tight if its pullback to every finite cover of $M$ remains tight, and virtually overtwisted if some finite cover becomes overtwisted; $\xi$ is Stein fillable if $(M,\xi)$ bounds a Stein domain (a complex-affine filling), which implies tightness. A rational homology 3-sphere is an L-space if its Heegaard Floer homology (an invariant from genus-one Heegaard diagrams) is as simple as that of a lens space; L-spaces admit no coorientable taut foliations (foliations by surfaces that a closed transverse loop meets on every leaf), and the broader L-space conjecture landscape expects them to be extremal for several tautness-type notions. Before 2023, classifications of tight contact structures existed for lens spaces, many Seifert fibered spaces, and torus bundles, but not for any infinite family of hyperbolic L-spaces.

Min and Nonino study $M(n,r)$, the result of $(n,r)$-surgery on the two components of the Whitehead link (a symmetric two-component hyperbolic link in $S^3$ with linking number zero); for infinitely many parameters these manifolds are hyperbolic L-spaces. For $0<r\le1$ write the negative continued fraction $-\frac{1}{r}=[r_0,\dots,r_n]$ with $r_0\le-1$ and $r_i\le-2$; they define $\Phi(r)=|r_0(r_1+1)\cdots(r_n+1)|$, extend it to all of $\mathbb{Q}$ by $\Phi(r+1)=\Phi(r)$, and set $\Psi(r)=0$ for $r=1$ and $\Psi(r)=\Phi(\frac{1}{1-r})$ otherwise; $\mathbb{Q}_-$ denotes the negative rationals and $\mathcal{R}_+=([2,4)\cup[5,\infty))\cap\mathbb{Q}$. Their main theorem: for every integer $n\ge5$ and $r\in\mathcal{R}_+\cup\mathbb{Q}_-$, the manifold $M(n,r)$ carries exactly $\Psi(r)+2\Phi(r)$ (for $r\in\mathcal{R}_+$) or $\Psi(r)$ (for $r\in\mathbb{Q}_-$) tight contact structures up to isotopy, pairwise distinguished by their contact invariants in Heegaard Floer homology (a homological invariant of a contact structure, used to separate structures). They also prove every one of these structures is Stein fillable (their Theorem 1.6) and at least $\Psi(r)+2\Phi(r)-6$ (resp. $\Psi(r)-2$) of them are virtually overtwisted (Theorem 1.7). The upper bound comes from a convex decomposition — a splitting of the manifold along convex surfaces (surfaces for which the contact structure is invariant under a transverse vector field, so that it is encoded by a dividing curve system) — of the complement of a genus-one fibered knot in the lens space $L(n,n-1)$, a once-punctured torus bundle over $S^1$ whose monodromy $\phi$ (the gluing homeomorphism of the fiber) is pseudo-Anosov (the exponential-stretch Thurston type of surface homeomorphism); they generalize this technique to monodromies whose fractional Dehn twist coefficient $c(\phi)$ (a rational measure of the boundary twist of the monodromy) satisfies $-1<c(\phi)<1$.

Two boundaries of these results motivate the present problem. First, by their Remark 1.3, for $n>2$ and $r\ge1$ (resp. $r<1$) at least $\Psi(r)+2\Phi(r)$ (resp. $\Psi(r)$) tight structures can still be constructed when $n<5$ or $r\notin\mathcal{R}_+$ (resp. $r\notin\mathbb{Q}_-$), but the authors warn "there could be more": no upper bound, hence no classification, is known there. The one exception is the slope $r=0$: $M(n,0)$ is a torus bundle over $S^1$, their Corollary 1.4 gives the exact count of tight contact structures on $M(n,m)$ for every integer $m\neq1,4$ when $n\ge5$ — in particular, infinitely many for $m=0$ — and the Colin–Giroux–Honda theorem quoted in the paper (a closed orientable irreducible 3-manifold supports infinitely many tight contact structures up to isotopy if and only if it is toroidal) settles $r=0$ for the remaining $n$ as well. Second, by their Remark 1.8, when the count is small (e.g. integer slopes $r\ge2$ with $\Psi(r)+2\Phi(r)<6$) the virtual-overtwistedness lower bound of Theorem 1.7 is vacuous, and it is unknown whether any tight structure on $M(n,r)$ is universally tight or whether all are virtually overtwisted. The paper also posed Conjecture 1.1 — a hyperbolic 3-manifold is an L-space if and only if it supports no universally tight contact structure — whose L-space-implies-no-universally-tight direction Baldwin, Sivek, and Zung refuted in 2025 by constructing hyperbolic L-spaces (large surgeries on hyperbolic L-space knots) that carry universally tight contact structures. Their examples do not lie in the Whitehead link family, so the universal-tightness question for the manifolds $M(n,r)$ themselves survives, now decoupled from the refuted conjecture.

Problem Statement

Let $M(n,r)$ denote $(n,r)$-surgery on the Whitehead link, $n\in\mathbb{Z}$, $r\in\mathbb{Q}$, and let $\Phi$, $\Psi$, $\mathcal{R}_+$, $\mathbb{Q}_-$ be as defined by Min and Nonino in the background. Resolve the two remaining threads of their classification program:

(1) Classification outside the resolved range: for integers $2<n<5$ (any rational $r\neq0$), and for $n\ge5$ with $r\in\mathbb{Q}\setminus(\mathcal{R}_+\cup\mathbb{Q}_-\cup\{0\})$ (including positive rationals in $[1,2)$ and $[4,5)$ and $0<r<1$), determine the exact number of tight contact structures on $M(n,r)$ up to isotopy and classify them. The slope $r=0$ is excluded because $M(n,0)$ is a torus bundle over $S^1$ whose count is already settled in the source: infinitely many tight contact structures up to isotopy (Corollary 1.4 for $n\ge5$, and the Colin–Giroux–Honda theorem cited there for the remaining $n$). In particular, decide whether the known lower bounds of $\Psi(r)+2\Phi(r)$ structures for $r\ge1$ (resp. $\Psi(r)$ for $r<1$) are sharp, or whether, as the authors warn, "there could be more".

(2) Universal tightness on the family: for each $M(n,r)$ in the classified range $n\ge5$, $r\in\mathcal{R}_+\cup\mathbb{Q}_-$ and in the ranges of (1), determine which of its tight contact structures are universally tight and which are virtually overtwisted. This includes the small-count regimes (e.g. integer slopes $r\ge2$ with $\Psi(r)+2\Phi(r)<6$) where currently no tight structure on $M(n,r)$ is known to be either universally tight or virtually overtwisted.

The completion boundary is the isotopy classification together with the universally-tight/virtually-overtwisted status of every tight structure on each $M(n,r)$ in the stated ranges. The source's broader expectation that its convex-decomposition technique extend to further pseudo-Anosov monodromies is context, not a deliverable.

References

1. Hyunki Min and Isacco Nonino, "Tight contact structures on a family of hyperbolic L-spaces," Geometriae Dedicata 220 (2026), DOI: 10.1007/s10711-026-01097-8; preprint arXiv:2311.14103, https://arxiv.org/abs/2311.14103
2. John A. Baldwin, Steven Sivek, and Jonathan Zung, "Pseudo-Anosov flows on hyperbolic L-spaces," arXiv:2505.21113 (2025), https://arxiv.org/abs/2505.21113
3. Mahan Mj and Balarka Sen, "Tight contact structures on hyperbolic homology 3-spheres," arXiv:2404.14698 (2024), https://arxiv.org/abs/2404.14698
4. John B. Etnyre, "Tight and overtwisted contact structures," arXiv:2504.13096 (2025), https://arxiv.org/abs/2504.13096
5. Hyunki Min and Konstantinos Varvarezos, "Bordered contact invariants and half Giroux torsion," arXiv:2506.14050 (2025), https://arxiv.org/abs/2506.14050
6. Vitalijs Brejevs and Andy Wand, "Stein-fillability and positivity in the mapping class group," arXiv:2405.07707 (2024), https://arxiv.org/abs/2405.07707