Shake genus and slice genus (Q2331024)

A central question within \(4\)-manifold topology is to determine the least genus of an embedded surface representing an arbitrary given class in \(H_2(M;\mathbb{Z})\) for a \(4\)-manifold \(M\). This is interesting even for the simple case of the \(n\)-trace of a knot \(K\), denoted \(X_n(K)\), namely the smooth manifold produced by attaching an \(n\)-framed \(2\)-handle to \(B^4\) along the knot \(K\subseteq S^3\), after smoothing corners. Note that \(H_2(X_n(K);\mathbb{Z})\cong \mathbb{Z}\), and the least genus of an embedded surface representing a generator of this group is called the \(n\)-shake genus of \(K\). A long-standing question on Kirby's celebrated problem list asked whether the \(0\)-shake genus of a knot must equal its slice genus. This paper resolves this question, by constructing infinitely many knots whose \(0\)-shake genus is strictly lower than the slice genus. The constructive part of the proof consists of a reinterpretation of a classical technique of Lickorish and Brakes, involving ``dualizable patterns'', to produce pairs of knots with diffeomorphic \(0\)-traces. A computation of Rasmussen's \(s\)-invariant then provides a lower bound on the slice genus. As a corollary, the author constructs bijective satellite maps on the smooth knot concordance group which fix the identity, but do not preserve slice genus. The question of whether there exist knots with trivial \(0\)-shake genus but nontrivial slice genus remains open.