Finite knot surgeries and Heegaard Floer homology (Q2346717)

One of the results that the paper under review proves is that if \(Y\) is a non-lens space elliptic manifold obtained by surgery on a knot \(K \subset S^3\) and \(H_1(Y) \leq 4\), then \(Y\) is one of the four spaces given in the paper and \(K\) is the trefoil. \textit{P. Ghiggini} [Am. J. Math. 130, No. 5, 1151--1169 (2008; Zbl 1149.57019)] proved that the Poincaré homology sphere can only be obtained by a surgery on the trefoil. The Poincaré homology sphere is the only elliptic homology sphere (other than \(S^3\)), so the result above can be seen as an extension of Ghiggini's result. The paper also lists all elliptic manifolds \(Y\) that are not lens spaces, have \(|H_1(Y)| \leq 32\) and can be obtained by surgery on a knot in \(S^3\). Moreover, if there is a knot that gives such a \(Y\) by surgery, then its Alexander polynomial coincides with the Alexander polynomial of a knot that is already known to have such a surgery. Elliptic manifolds are \(L\)-spaces, i.e. `have as simple Heegaard Floer homology as possible'. Their Heegaard Floer homology can be completely described by a set of numbers, called \textit{correction terms}. If an \(L\)-space \(Y\) is obtained by \(p/q\) surgery on a knot \(K\) there is a formula that expresses the correction terms of \(Y\) in terms of the Alexander polynomial of \(K\) and \(p/q\). Moreover, there is an upper bound on the genus of \(K\), \(K\) must be fibred and its Alexander polynomial takes a very specific form. On the other hand, Heegaard Floer homology of elliptic manifolds (and other rational homology sphere Seifert fibred spaces) can be described algorithmically using plumbing graphs [\textit{P. Ozsváth} and \textit{Z. Szabó}, Geom. Topol. 7, 185--224 (2003; Zbl 1130.57302)]. The main idea of the paper under review is to compare the two descriptions of correction terms and thus to rule out some Alexander polynomials and slopes \(p/q\) (and so manifolds as well).