Invariants of boundary link cobordism

From MaRDI portal
Publication:4432912



Abstract: An n-dimensional mu-component boundary link is a codimension 2 embedding of spheres L=�igsqcup_{mu}S^n subset S^{n+2} such that there exist mu disjoint oriented embedded (n+1)-manifolds which span the components of L. An F_mu-link is a boundary link together with a cobordism class of such spanning manifolds. The F_mu-link cobordism group C_n(F_mu) is known to be trivial when n is even but not finitely generated when n is odd. Our main result is an algorithm to decide whether two odd-dimensional F_mu-links represent the same cobordism class in C_{2q-1}(F_mu) assuming q>1. We proceed to compute the isomorphism class of C_{2q-1}(F_mu), generalizing Levine's computation of the knot cobordism group C_{2q-1}(F_1). Our starting point is the algebraic formulation of Levine, Ko and Mio who identify C_{2q-1}(F_mu) with a surgery obstruction group, the Witt group G^{(-1)^q,mu}(Z) of mu-component Seifert matrices. We obtain a complete set of torsion-free invariants by passing from integer coefficients to complex coefficients and by applying the algebraic machinery of Quebbemann, Scharlau and Schulte. Signatures correspond to `algebraically integral' simple self-dual representations of a certain quiver (directed graph with loops). These representations, in turn, correspond to algebraic integers on an infinite disjoint union of real affine varieties. To distinguish torsion classes, we consider rational coefficients in place of complex coefficients, expressing G^{(-1)^q,mu}(Q) as an infinite direct sum of Witt groups of finite-dimensional division Q-algebras with involution. Numerical invariants of such Witt groups are available in the literature.


A {link} is a collection of \(n\)-dimensional spheres \(S^n\) disjointly embedded in \(S^{n+2}\). A {boundary} link is one whose components are the boundaries of disjoint \((n+1)\)-dimensional submanifolds of \(S^{n+2}\). This monograph concerns the classification of \(F_\mu\)-links, a refinement of the notion of a \(\mu\)-component boundary link, for \(n\geq 2\). The cobordism group of such links is denoted \(C_n(F_\mu)\). The first main result (Theorem A) gives an expression for \(C_{2q-1}(F_\mu)\) as a direct sum of free abelian groups and torsion groups \(\mathbb Z/2\), \(\mathbb Z/4\) and \(\mathbb Z/8\) (countably many summands of each), for \(\mu \geq 2\) and \(2q-1>1\). This is a generalization of the result of Levine for \(\mu=1\). The second main result (Theorem B, p. 26) gives a list of explicit invariants which decide when two \(2q-1\)-dimensional \(\mu\)-component links are cobordant. The proofs of these theorems are algebraic, and the starting point is the description of the link cobordism groups \(C_{2q-1}(F_\mu)\) by \textit{K. H. Ko} [Trans. Am. Math. Soc. 299, 657--681 (1987; Zbl 0626.57013)] and \textit{W. Mio} [Math. Proc. Camb. Philos. Soc. 101, 259--266 (1987; Zbl 0632.57015)] in terms of a Witt group of Seifert forms. These groups are analysed by means of a localisation sequence, from which invariants in the Witt groups of finite dimensional division algebras (over \(\mathbb Q\)) are obtained. This is the source of the explicit invariants used in Theorem B, such as signatures and Hasse-Witt invariants. In order to carry out the details, the author explains and uses a lot of interesting material concerning representation varieties, quivers, real algebraic sets, hermitian Morita theory, orderings and signatures.











This page was built for publication: Invariants of boundary link cobordism

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q4432912)