Dissolving four-manifolds and positive scalar curvature (Q1434510)

This paper has two separate parts. The first part covers the geography of symplectic 4-manifolds that dissolve with the addition of a single \(S^2\times S^2\). The second part shows that there are infinitely many smooth 4-manifolds that do not admit metrics of positive curvature with odd order fundamental group and universal cover with positive scalar curvature. This provides a counter-example to a conjecture of \textit{J.~Rosenberg} which was probably intended for dimensions five and higher [Topology 25, 319--336 (1986; Zbl 0605.53020)]. The notion of dissolving 4-manifolds has roots in the stabilization result of Wall stating that simply connected homotopy equivalent smooth 4-manifolds become diffeomorphic after connected sum with enough copies of \(S^2\times S^2\). Later \textit{R.~ Mandelbaum} gave a nice description of the \(4+1\) dimensional cobordism relating a fiber sum to its factors. He used this to study the behavior of algebraic surfaces under connected sum with \(\mathbb C P^2\) [\textit{R. Mandelbaum}, Decomposing analytic surfaces. In: Geometric Topology, Proc. Conf., Athens/Ga. 1977, 147--217 (1979; Zbl 0507.57011)]. \textit{R.~Gompf} introduced the notion of dissolving 4-manifolds and compiled the most important existence results for them [J. Differ. Geom. 34, 93--114 (1991; Zbl 0751.14024)]. A 4-manifold is said to dissolve if it becomes diffeomorphic to a connected sum of \(\mathbb C P^2\)'s and \(\overline{\mathbb C P^2}\)'s or a connected sum of \(K3\)'s and \(S^2\times S^2\)'s after one stabilization. The first main result in this paper fills in a large portion of the geography of symplectic 4-manifolds with symplectic 4-manifolds that dissolve. The authors modify \textit{J.~Park}'s constructions from [Math. Z. 240, 405--421 (2002; Zbl 1030.57032)] in order to smoothly apply the results of Gompf. The main application of dissolving symplectic 4-manifolds is to the behavior of positive scalar curvature under finite covers. This is the second part of the paper. It is known that connected sums of \(S^2\times S^2\)'s admit metrics of positive scalar curvature and that a metric of positive scalar curvature together with some mild topological assumptions implies that the Seiberg-Witten invariants vanish. Finally, any symplectic 4-manifold has non-trivial Seiberg-Witten invariants by the work of C.~Taubes. The construction of the main family of examples starts with a 3-dimensional spherical space form times \(S^1\). The generator of the fundamental group corresponding to \(S^1\) may be killed by a surgury and the resulting manifold may be added to one of the dissolving symplectic manifolds from the first part of the paper. The Seiberg-Witten invariants of any of the resulting manifolds are non-trivial by a gluing theorem from [\textit{D. Kotschick, J. W. Morgan} and \textit{C. H. Taubes}, Math. Res. Lett. 2, 119--124 (1995; Zbl 0853.57020)]. However, the universal cover of each of these manifolds dissolves by the results from the first part of this paper. Thus these manifolds do not admit metrics of positive scalar curvature even though their universal covers do. The authors do a good job stating all of the required results so the reader does not have to search for background material. The authors also include a number of interesting remarks and related results that nicely complete the paper.