On definite strongly quasipositive links and L-space branched covers (Q2335483): Difference between revisions

A strongly quasipositive link is a link which is a boundary of a quasipositive Seifert surface, a surface obtained from \(n\) parallel disks by attaching positive bands. To make this precise let \(B_n\) be the braid group on \(n\) strands with generators \(\sigma_1, \dots, \sigma_{n-1}\), and let \(a_{rs} = (\sigma_r \sigma_{r+1}\cdots \sigma_{s-2})\sigma_{s-1}(\sigma_r \sigma_{r+1}\cdots \sigma_{s-2})^{-1}\), where \(1 \le r < s \le n\). Then a strongly quasipositive link is realized as the closure of a braid of the form \(P = \prod_{k= 1}^m a_{r_k s_k}\), which is called a strongly quasipositive braid. A rational homology \(3\)-sphere \(M\) is an L-space if \(\dim\widehat{HF}(M; \mathbb{Z}_2) = | H_1(M; \mathbb{Z})|\), and a knot \(K\) in \(S^3\) is an L-space knot if it admits a nontrivial surgery yielding an L-space. An L-space knot is fibred [\textit{Y. Ni}, Invent. Math. 170, No. 3, 577--608 (2007; Zbl 1138.57031)], strongly quasipositive [\textit{M. Hedden}, J. Knot Theory Ramifications 19, No. 5, 617--629 (2010; Zbl 1195.57029)] and prime [\textit{D. Krcatovich}, Topology Appl. 194, 171--201 (2015; Zbl 1326.57020)]. In the paper under review, the authors address the question: for which fibred strongly quasipositive links \(L\) is some \(\Sigma_n(L)\), the canonical \(n\)-fold cyclic branched cover of \(S^3\) branched over \(L\), an L-space? Let \(L\) be a strongly quasipositive link. Then it is a closure of a strongly quasipositive braid \(\delta_n^k P\), where \(n \ge 2,\ k \ge 0\), for some strongly quasipositive braid \(P \in B_n\) and the dual Garside element \(\delta_n = \sigma_1 \sigma_2 \dots \sigma_{n-1} \in B_n\). We call the maximum of such \(k \ge 0\) the BKL-exponent of \(L\), and denote it by \(k(L)\). We say that \(L\) is definite if \(| \sigma(L)| = 2g(L) + (m-1)\), where \(\sigma(L)\) is the signature of \(L\), \(g(L)\) is the genus of \(L\) and \(m\) denotes the number of components of \(L\). Then for a prime strongly quasipositive link \(L\) the authors prove that \(L\) is definite and \(k(L) \ge 2\) if and only if it is one of the torus links \(T(2, q), T(3, 4), T(3, 5)\) or pretzel links \(P(-2, 2, m), P(-2, 3, 4)\). For such a link \(L\), \(\Sigma_2(L)\) has finite fundamental group and is therefore an L-space [\textit{P. Ozsváth} and \textit{Z. Szabó}, Topology 44, No. 6, 1281--1300 (2005; Zbl 1077.57012)]. Applying the above result, together with a previous result of the authors [\textit{M. Boileau} et al., J. Topol. 12, No. 2, 536--576 (2019; Zbl 1422.57012)], they prove that for a prime strongly quasipositive link \(L\) with \(k(L) \ge 2\), \(\Sigma_n(L)\) is an L-space for some \(n \ge 2\) if and only if \(L\) is one of the above links. (Note that a strongly quasipositive link \(L\) with \(k(L) \ge 1\) is fibred.) The assumption \(k(L) \ge 2\) in the result leads them to conjecture that if \(L\) is a prime, fibred, strongly quasipositive link with BKL-exponent \(k(L) \le 1\), then no \(\Sigma_n(L)\) is an L-space. Furthermore, the authors prove that if \(K\) is a non-trivial L-space knot of braid index 2 or 3 and some \(\Sigma_n(L)\) is an L-space, then \(K\) is one of the torus knots \(T(2, m)\) where \(m \ge 3\) is odd, \(T(3,4)\), or \(T(3, 5)\).