Symmetric chain complexes, twisted Blanchfield pairings and knot concordance (Q1631704): Difference between revisions

The paper under review is concerned with linking forms and their applications to knot concordance. Given an integral domain \(R\) with involution and field of fractions \(Q\), a linking form over \(R\) is a sesquilinear Hermitian pairing \(H \times H \to Q/R\) on a torsion \(R\)-module \(H\). For instance, given a 3-manifold \(M\), the torsion submodule of \(H_1(M;\mathbb{Z})\) supports a linking form over \(\mathbb{Z}\), while the Alexander module \(H_1(S^3 \setminus \nu K;\mathbb{Z}[t^{\pm 1}])\) of a knot \(K \subset S^3\) supports a linking form over \(\mathbb{Z}[t^{\pm 1}]\) called the Blanchfield form. These constructions can be generalized to various twisted coefficients, cf. [\textit{T. D. Cochran} et al., Ann. Math. (2) 157, No. 2, 433--519 (2003; Zbl 1044.57001), \textit{C. Leidy}, Comment. Math. Helv. 81, No. 4, 755--781 (2006; Zbl 1105.57003), \textit{S. Friedl} et al., Homology Homotopy Appl. 19, No. 2, 275--287 (2017; Zbl 1401.57012)]. We now describe the structure of this paper. The first part recalls a chain complex description of twisted Blanchfield pairings from [\textit{M. Powell}, Q. J. Math. 67, No. 4, 715--742 (2016; Zbl 1365.57009)]. The second part provides an explicit algorithm to compute twisted Blanchfield pairings, see Algorithm 1.2. The third part is concerned with applications to knot concordance. To describe these latter results in more detail, we recall some relevant definitions. A knot \(K \subset S^3\) is called topologically slice if it bounds a locally flat disc in the 4-ball \(D^4\). Slice knots are algebraically slice (i.e. admit metabolic Seifert matrices), but the converse is not true as demonstrated by Casson and Gordon [\textit{L. Guillou} (ed.) and \textit{A. Marin} (ed.), A la recherche de la topologie perdue. 1: Du côté de chez Rohlin. 2: Le côté de Casson. Birkhäuser/Springer, Basel (1986; Zbl 0597.57001)]. Later work related to the Casson-Gordon invariants includes Kirk and Livingston's study of twisted Alexander polynomials [\textit{P. Kirk} and \textit{C. Livingston}, Topology 38, No. 3, 635--661 (1999; Zbl 0928.57005)], as well as the solvable filtration introduced by Cochran, Orr and Teichner [\textit{T. D. Cochran} et al., Ann. Math. (2) 157, No. 2, 433--519 (2003; Zbl 1044.57001)]. In Theorem 1.3, the authors show that if a knot is slice, then certain metabelian Blanchfield pairings are metabolic. In particular, this recovers an obstruction due to Kirk and Livingston. An obstruction to slicing satellite knots is stated in Theorem 1.4; it is based on ideas of [\textit{T. D. Cochran} et al., Geom. Topol. 13, No. 3, 1419--1482 (2009; Zbl 1175.57004)]. The paper concludes with several enlightening examples: for instance Example 8.2 applies Theorem 1.4 to obstruct the sliceness of a knot with vanishing Casson-Gordon invariants; the proof is simpler than the one originally given in Section 6 of [\textit{T. D. Cochran} et al., Ann. Math. (2) 157, No. 2, 433--519 (2003; Zbl 1044.57001)].