Every knot has characterising slopes (Q2423419)

The author proves that every knot \(K\subset S^3\) has infinitely many characterising slopes. Here a slope \(p/q\) is said to be a characterising slope of a knot \(K\subset S^3\) if whenever the Dehn filling \(S^3_{K}(p/q)\) is homeomorphic to the Dehn filling \(S^3_{K'}(p/q)\) of another knot \(K'\subset S^3\), via an orientation-preserving homeomorphism, then \(K\) and \(K'\) are isotopic to each other. More precisely, the author proves that for any knot \(K\subset S^3\), \(p/q\) is a characterising slope if \(|p|<|q|\) and \(|q|\) is greater than a constant only depending on \(K\). In previous works, Kronheimer, Mrowka, Ozsváth and Szabó proved that each slope of the unknot is a characterising slope [\textit{P. B. Kronheimer} et al., Ann. Math. (2) 165, No. 2, 457--546 (2007; Zbl 1204.57038)]; and \textit{D. McCoy} proved that for each torus knot, all but finitely many non-integral slopes are its characterising slopes [``Non-integer characterizing slopes for torus knots'', Preprint, \url{arXiv:1610.03283}]. In this paper, the author proves the main result by first proving it for hyperbolic knots, then taking care of the satellite knot case (i.e. the knot complement has nontrivial JSJ decomposition). In the case that both \(K\) and \(K'\) are hyperbolic knots, the main result basically follows from Thurston's hyperbolic surgery theorem and a proof by contradiction. By assuming that there is a sequence of knots \(K_i\subset S^3\) and slopes \(p_i/q_i\) such that \(S^3_K(p/q)\cong S^3_{K_i}(p_i/q_i)\) and \(|q_i|\to \infty\), the author proves that, up to taking a subsequence, all \(S^3\setminus N(K_i)\) are Dehn fillings of one single cusped hyperbolic \(3\)-manifold \(X\). Then the shortest closed geodesics in the sequence \(\{S^3_{K_i}(p_i/q_i)\}\) must have length going to \(0\), which contradicts that all of these manifolds are homeomorphic to \(S^3_K(p/q)\). In the case that \(K\) is a satellite knot, if \(|p|<|q|\) and \(|q|>9\), then the author proves that the decomposition tori of \(S^3\setminus N(K)\) also give the torus decomposition of \(S^3_K(p/q)\). This result relies heavily on Budney's classification of all possible JSJ pieces of knot complements in \(S^3\) [\textit{R. Budney}, Enseign. Math. (2) 52, No. 3--4, 319--359 (2006; Zbl 1114.57004)]. When \(|q|\) is large enough, the author further proves that the filling core is either the shortest closed geodesic in all hyperbolic pieces of \(S^3_K(p/q)\), or a singular fiber of the largest index in all Seifert pieces of \(S^3_K(p/q)\). This property basically guarantees that the homeomorphism between \(S^3_K(p/q)\) and \(S^3_{K'}(p/q)\) must map the filling core of \(S^3_K(p/q)\) to the core of \(S^3_{K'}(p/q)\). So the two knots \(K\) and \(K'\) are isotopic to each other.