The Euler characteristics of \(SU(3)\)-instanton and moduli spaces over four sphere (Q5961481)

Let \({\mathfrak M}_k (n)\) denote the moduli space of \(SU(n)\)-instanton bundles over the 4-sphere \(S^4\) with second Chern class \(k\in \mathbb{Z}\). These moduli spaces play a central role in the geometric theory of Yang-Mills fields, and over the past twenty years they have been objects of intensive studies in complex geometry and its related topology. Despite the great progress achieved in describing the topology of those moduli spaces, which is mainly provided by the celebrated Donaldson theory [cf. \textit{S. K. Donaldson}, Commun. Math. Phys. 93, 453-461 (1984; Zbl 0581.14008)] and the recent work done around the Atiyah-Jones conjecture [cf. \textit{C. P. Boyer}, \textit{J. C. Hurtubise}, \textit{B. M. Mann} and \textit{R. J. Milgram}, Ann. Math., II. Ser. 137, No. 3, 561-609 (1993; Zbl 0816.55002)], there are still some crucial open questions to be answered. One of them, which is also of particular interest in mathematical physics [cf. \textit{C. Vafa} and \textit{E. Witten}, Nucl. Phys. B 431, No. 1-2, 3-77 (1994)], is given by the problem of computing the topological Euler characteristics of the moduli spaces \({\mathfrak M}_k (n)\) for \(k>1\).NEWLINENEWLINENEWLINEIn the present note, the author develops a method for calculating the Euler characteristics \(\chi ({\mathfrak M}_k (2))\) and \(\chi ({\mathfrak M}_k (3))\) explicitly as number-theoretic functions of \(k>1\). This method is based upon the study of the circle action on the two moduli spaces under investigation, with a concrete description of the corresponding fixed point sets. As for the first case \((n=2)\) considered here, the identity \(\chi ({\mathfrak M}_k (2))= E(k):=\) ``number of divisors of \(k\)'' had been derived, in a different way, by \textit{M. Furuta} [cf. Proc. Japan Acad., Ser. A 63, 266-267 (1987; Zbl 0645.58044)]. However, the author's method does not only provide the precise value of \(\chi ({\mathfrak M}_k (3))\), but also might be the right strategy for computing the Euler characteristics \(\chi ({\mathfrak M}_k (n))\) in the general, apparently much more involved case of arbitrary \(n\geq 4\).