An obstruction to slicing knots using the eta invariant (Q4954265)
From MaRDI portal
!
This is the item page for this Wikibase entity, intended for internal use and editing purposes. Please use the normal view instead:
scientific article; zbMATH DE number 1449187
| Language | Label | Description | Also known as |
|---|---|---|---|
| default for all languages | No label defined |
||
| English | An obstruction to slicing knots using the eta invariant |
scientific article; zbMATH DE number 1449187 |
Statements
An obstruction to slicing knots using the eta invariant (English)
0 references
22 March 2001
0 references
slice knots
0 references
eta invariant
0 references
knot concordance
0 references
Casson-Gordon invariants
0 references
0.8038875460624695
0 references
0.8011112213134766
0 references
0.772819995880127
0 references
0.7709967494010925
0 references
0.7594869136810303
0 references
Fix a group \(G\) and consider manifolds together with homomorphisms of their fundamental groups to \(G\). Two \(n\)-manifolds are called homology \(G\)-bordant if they cobound an \((n+1)\)-manifold with a homomorphism of its fundamental group to \(G\) and such that this manifold has the homology type of a product. In 1989, the reviewer showed that the reduced eta invariant associated to a finite dimensional unitary representation could be used to detect links not concordant to boundary links, a phenomenon first exhibited by Cochran-Orr. However this technique gave no new information about knot concordance. In the present paper the author constructs a new group \(G\) aimed specifically at the knot concordance problem, allowing him to define secondary knot concordance invariants, akin to the Casson-Gordon invariants. This group is a semidirect product of a torsion \(\mathbb Z [\mathbb Z]\)-module and the integers \(\mathbb Z\), and is the fundamental new contribution of this work. Unlike the link problem, but similar to the knot concordance obstructions of Casson and Gordon, homomorphisms to this ``Letsche group'' are determined not by the knot group alone, but by additional structure -- a choice of element in the Alexander module together with Poincaré duality (in this case, via the Blanchfield form of the knot.) The main result is the following. If a knot \(K\) is slice, then there is a submodule, \(P\), of the Alexander module such that \(P = P^{\perp} = \{ y \;|\;B(x,y) = 0 \text{ for every } x \in P\}\), and such that for every \(x \in P\) such that the corresponding unitary representation is outside a ``special subvariety'' of the representation ring of the Letsche group, the eta invariant vanishes. Here, \(B\) is the Blanchfield pairing of the knot. Recognizable criteria are given guaranteeing that a given representation lies outside the special subvariety (Theorem 3.11), and in contrast to the Casson and Gordon invariants, representations of finite non-prime power covers that factor through non-cyclic \(p\)-groups satisfy the Letsche criteria. Sadly, no examples for specific knots are given in the paper, and the relation to Casson-Gordon invariants remains open.
0 references