Branched covers bounding rational homology balls (Q2070104)

Given a knot \(K\) in \(S^3\) and \(\Sigma_q(K)\) the \(q\)-fold cyclic branched covering of \(S^3\) along \(K\), where \(q\) is a prime power, it is known that \(\Sigma_q(M)\) is a rational homology sphere. Moreover, in the case that \(K\) bounds a smooth, properly embedded disk in the four dimensional ball, i.e., \(K\) is slice, then \(\Sigma_q(K)\) bounds a smooth rational homology ball. This is not a general behaviour for slice knots. The authors give examples of non-slice knots satisfying the mentioned property. They use twisted Alexander polynomials to show that the given knots are not slice: ``Twisted Alexander polynomials are associated to a choice of \(q\in \mathbb{Q}\) and a map \(\chi:H_1(\Sigma_q(K); \mathbb{Z}) \rightarrow \mathbb{Z}_d\) for some \(d\). In order to use twisted Alexander polynomials to obstruct a knot \(K\) from being slice, one must show that for every subgroup \(M\) of \(H_1(\Sigma_q(K); \mathbb{Z})\) satisfying certain algebraic properties, there exists a map \(\chi\) vanishing on \(M\) such that the resulting twisted Alexander polynomial does not factor in a certain way.''