The Dolgachev surface. Disproving the Harer-Kas-Kirby conjecture. (Q662444): Difference between revisions

The Dolgachev surface \(E(1)_{2,3}\) is an elliptic complex surface which is obtained from the standard \(E(1) = \mathbb{CP}^2 \sharp \overline{\mathbb{CP}}^2\) by performing two logarithmic transformation operations of order \(2\) and \(3\) on two disjoint fibers. It is also known that the Dolgachev surface can be obtained from \(E(1)\) by performing knot surgery by using the trefoil knot \(K\) [\textit{J. Park}, Bull. Lond. Math. Soc. 36, No. 2, 231--240 (2004; Zbl 1064.57029)]. The Harer-Kas--Kirby conjecture states that the Dolgachev surface \(E(1)_{2,3}\) must contain \(1\)- and \(3\)-handles [\textit{J. Harer, A. Kas} and \textit{R. Kirby}, Handlebody decompositions of complex surfaces. Mem. Am. Math. Soc. 350, 102 p. (1986; Zbl 0631.57009)]. \textit{K. Yasui} constructed an exotic \(\mathbb{CP}^2 \sharp \overline{\mathbb{CP}}^2\) which shares the same Seiberg-Witten invariants with \(E(1)_{2,3}\) but which has no \(1\)- and \(3\)-handles [Algebr. Geom. Topol. 8, 971--996 (2008; Zbl 1161.57017)]. In the article under review, the author shows that the Dolgachev surface \(E(1)_{2, 3}\) also admits a handlebody structure without \(1\)- and \(3\)-handles so the Harer-Kas-Kirby conjecture is disproved. The main tool of the proof is Kirby calculus including slice \(1\)-handles, \(1\)-handle/\(2\)-handle pair cancellations and turning a handlebody upside down. The author also locates a ``cork'' inside of \(E(1)_{2,3}\), so that \(E(1)_{2,3}\) is obtained from \(E(1)\) by twisting along this cork. This result is extended in [\textit{S. Akbulut}, J. Gökova Geom. Topol. GGT 3, 22-43 (2009; Zbl 1267.57036)].