A decomposition of smooth simply-connected \(h\)-cobordant 4-manifolds (Q679319)

Motivated by an example of \textit{S. Akbulut} [J. Differ. Geom. 33, No. 2, 335-356 (1991; Zbl 0839.57015)], where he gives an example of two \(h\)-cobordant but non-diffeomorphic 4-manifolds \(M,N,\) where \(N\) is formed from \(M\) by cutting out a contractible submanifold and regluing it via a nontrivial involution of the boundary, the author gives the following decomposition theorem for h-cobordant simply connected smooth 4-manifolds: Let \(U\) be a smooth 5-dimensional simply-connected h-cobordism with \(\partial U = M_1 \cup (-M_2).\) Let \(f: M_1 \to M_2\) be the homotopy equivalence induced by \(U.\) Then the following hold: (1) There are decompositions \(M_1 = M \#_\Sigma W_1, M_2 = M \#_\Sigma W_2 \) such that \(in_{2*} \circ in_{1*}^{-1} = f_*: H_2(M_1) \to H_2(M_2).\) Here \(in_{1*}, in_{2*}\) are the maps induced in the second homology by embeddings of \(M\) into \(M_1\) and \(M_2\) respectively, and \(W_1,W_2\) are smooth, compact, contractible 4-manifolds and \(\Sigma = \partial W_1 = \partial W_2 = \partial M;\) (2) These decompositions may be chosen so that \(W_1\) is diffeomorphic to \(W_2.\) The proof also shows that the whole cobordism can be decomposed into subcobordisms, one of which is a product cobordism and the other of which is diffeomorphic to \(D^5\) as a smooth manifold. The techniques involve handle manipulations and Kirby calculus. The main result, with slightly different refinements, was also proved independently by \textit{C. L. Curtis} and \textit{W. C. Hsiang} and simplified and refined with contributions from \textit{M. H. Freedman} and \textit{R. Stong}; this work was published jointly in [Invent. Math. 123, No. 2, 343-348 (1996; Zbl 0843.57020)]. One of their refinements was to show that one could choose \(M\) to be simply connected. Further refinements and an exposition of the problem were provided by \textit{R. Kirby} [Turk. J. Math. 20, No. 1, 85-93 (1996; Zbl 0868.57031)]. One of the additional refinements discussed there (and attributed to the paper under review) is that \(W_1\) is diffeomorphic to \(W_2\) via a diffeomorphism which is an involution on the boundary, as in Akbulut's original example. In particular, these imply that any smooth homotopy 4-sphere arises by cutting out a contractible 4-manifold from \(S^4\) and gluing it back by an involution of the boundary. Kirby also notes the delicate nature of the non-trivial h-cobordisms between contractible manifolds constructed this way in that they can be trivialized by adding \(W_1 \times I\) along \(\partial W_1 \times I.\)