Instantons, concordance, and Whitehead doubling (Q713192)

In this nice paper Hedden and Kirk show that the Whitehead doubles of the \((2,2^n-1)\) torus knots are independent in the smooth concordance group. Their proof relies on their earlier analysis of the \(\mathbb{Z}_2\) homology cobordism group via Yang-Mills gauge theory. They are careful to abstract the topological assumptions out of the gauge theory arguments in the hope that topologists who may be less familiar with gauge theory may be able to apply their techniques to other problems. In this vein they define property \(\mathcal{I}\) for certain negative definite \(4\)-manifolds to mean that the moduli of anti-self-dual \(\text{SO}(3)\) connections has positive dimension and an odd number of singular points. The singular points in this moduli space are given by reducible connections, and these are characterized homologically via Hodge theory. The dimension of the moduli space may be computed via the index theorem. They show that this property is preserved under certain gluing operations. To complete the gauge theory argument one must show that the associated moduli space is compact. The ways that these moduli spaces can fail to be compact are well understood. Energy can bubble off at points -- but only in fixed quanta, and energy can slide off the end. This last behavior is determined by the Chern-Simons invariants of the flat connections on the boundary components. After describing these issues in general, the authors analyze the \(2\)-fold branched cover of a Whitehead double of a knot. They show that it decomposes into a copy of the exterior of the \((2,4)\)-torus link and two copies of the exterior of the knot. They use this description to do two things -- construct a suitable \(4\)-manifold having the \(2\)-fold branched cover of the Whitehead double of a torus knot along with three lens spaces as boundary, and to analyze the representation space and estimate the Chern-Simons invariants of the \(2\)-fold branched cover of a Whitehead double of a torus knot. The paper ends with an insightful comparison of various techniques used to analyze concordance and argues that Yang-Mills gauge theory is still the only tool that is able to achieve the present results.