A gluing formula for families Seiberg-Witten invariants (Q2004528): Difference between revisions

Seiberg-Witten theory constructs invariants of smooth four-manifold out of moduli spaces of solutions of the Seiberg-Witten equation. The introduction of these invariants revolutionised gauge theory and low-dimensional topology in the 1990s. To this day Seiberg-Witten theory remains one of the crucial tools in low-dimensional topology. Donaldson proposed to study a version of these invariants for families of four-manifolds [\textit{S. K. Donaldson}, Bull. Am. Math. Soc., New Ser. 33, No. 1, 45--70 (1996; Zbl 0872.57023)]. This idea has been picked up by many authors. A particularly striking application of the families Seiberg-Witten invariant is Ruberman's proof that for certain simply connected 4-manifolds the space of positive scalar metrics has infinitely many connected components [\textit{D. Ruberman}, Geom. Topol. 5, 895--924 (2001; Zbl 1002.57064)]. To compute the families Seiberg-Witten invariant \textit{T.-J. Li} and \textit{A.-K. Liu} established a wall-crossing formula [Commun. Anal. Geom. 9, No. 4, 777--823 (2001; Zbl 1034.53090)] and \textit{A.-K. Liu} established a blow-up formula [J. Differ. Geom. 56, No. 3, 381--579 (2000; Zbl 1036.14014)]. The article under review establishes gluing formulae for the families Seiberg-Witten invariant -- a far-reaching generalisation of a crucial part of Ruberman's argument. The authors also discuss the consequences of their formulae in detail and provide a substantial number of topological applications. The formulations of the precise gluing formulae are rather involved but exceptionally well explained in the article (see Theorems 1.1, 4.1, 4.3, 4,4). Here are some of the applications: \begin{itemize} \item The four-manifolds \(X_n = (S^2 \times S^2)^{\# n} \,\#\, K3^{\# n}\) (\(n\geq 2\)) and \(Y_{n,m} = (\mathbb{C}P^2)^{\# 2n} \,\#\, (\overline{\mathbb{C}P^2})^{\# m}\) (\(n\geq 2, m \geq 10n + 1\)) admit diffeomorphisms which are isotopic to the identity but not smoothly so. \item The four-manifold \(Y_{n,m}\) (\(n \geq 4\), \(m \geq n + 18\)) admits a metric of positive scalar curvature and smooth involution \(f\) but no \(f\)-invariant metric of positive scalar curvature. \end{itemize}