Primary decompositions of knot concordance (Q2045869)

A central problem in three-dimensional knot concordance theory in the smooth category consists of understanding \(\mathcal{T}\), the concordance group of topologically slice knots. In this paper, the author shows that primary decompositions of \(\mathcal{T}\) of a strong type cannot exist. The main goal of this paper is to provide a counterexample to a splitting property related to the surjectivity of \(\Phi\), considered by Cha under the name strong existence [\textit{J. C. Cha}, Forum Math. Sigma 9, Paper No e57, 37 p. (2021; Zbl 1473.57006)]. The main theorems are listed as follows. \textbf{Theorem 1.1} There exists a set of three polynomials , \(\mathcal{P}_0=\{f_1,f_2,f_3\}\subset \mathcal{P}\), such that the natural homomorphism \[ \mathcal{T}_{\Delta}^{f_1}\oplus \mathcal{T}_{\Delta}^{f_2}\oplus\mathcal{T}_{\Delta}^{f_3}\rightarrow \mathcal{T}_{\Delta}^{\mathcal{P}_0} \] is not surjective. \textbf{Theorem 1.2} There exist Alexander polynomials \(f_1(t)\) and \(f_2(t)\) having no common factors and a topologically slice knot \(K\) with \(\Delta_K(t)=f_1(t)^2f_2(t)^2\) such that \(K\) is not concordant to any connected sum of knots \(K_1\sharp K_2\) where \(\Delta_{K_i}(t)=f_i(t)^{n_i}\) and \(n_1,n_2\in \mathbb{Z}\). It follows that \(K\) is a topologically slice knot that is not smoothly concordant to a knot with Alexander polynomial one. The first examples of such knots were described in [\textit{M. Hedden} et al., Adv. Math. 231, No. 2, 913--939 (2012; Zbl 1254.57008)]. In the appendix of this paper, the author also describes the computation of the \(d\)-invariants for the 2-fold branched cover of \(S^3\) branched over \(K_{15}\).