Exotic smoothings via large \(\mathbb{R}^4\)'s in Stein surfaces (Q2630678)

The study of smoothing theory on open 4-manifolds is an important theme in differential topology. A reference point in this area is a 1982 paper by \textit{F. Quinn} [J. Differ. Geom. 17, 503--521 (1982; Zbl 0533.57009)] in which he shows that any open 4-manifold admits at least one smooth structure. It was later seen that there are many examples that admit uncountably many diffeomorphism classes of smooth structures, but it is still unknown if this can be expected in general. Methods for finding infinite families of exotic open 4-manifolds are based on information about the behavior of smoothings on \(\mathbb{R}^4\). \textit{R. E. Gompf} [J. Differ. Geom. 21, 283--300 (1985; Zbl 0562.57009)] defined an infinite collection of exotic \(\mathbb{R}^4\)'s, smooth 4-manifolds that are homeomorphic but not diffeomorphic to \(\mathbb{R}^4\). Then uncountably many exotic \(\mathbb{R}^4\)'s were produced by \textit{C. H. Taubes} [J. Differ. Geom. 25, 363--430 (1987; Zbl 0615.57009)]. This was the first 4-manifold to exhibit such uncountable behavior. Related families of smooth structures were subsequently defined on more general open 4-manifolds by fixing some standard smooth structure and then attaching infinitely many different \(\mathbb{R}^4\)'s using an operation introduced by \textit{R. E. Gompf} [J. Differ. Geom. 18, 317--328 (1983; Zbl 0496.57007)] called \textit{end-summing}. In this way, by some conditions on the same underlying 4-manifold infinitely many diffeomorphism classes (countable or uncountable) are obtained. In most cases, the resulting diffeomorphism types were distinguishable either by the \textit{Taylor invariant} [\textit{L. R. Taylor}, Geom. Topol. 1, 71--89 (1997; Zbl 0885.57012)] (taking values in \(\mathbb{Z}^{\geq 0}\cup \{\pm\infty\}\)) or a generalization of \textit{compact} \textit{equivalence classes} as in Gompf [1985, loc. cit.], which are two invariants that measure the complexity of embedded exotic \(\mathbb{R}^4\)'s in a given smooth 4-manifold. Finally, recently Gompf used \textit{Stein surfaces} to study open 4-manifolds [\textit{R. E. Gompf}, ``Minimal genera of open 4-manifolds'', Preprint, \url{arXiv:1309.0466}]. A Stein surface is an open complex 4-manifold that admits a proper biholomorphic embedding into some ambient \(\mathbb{C}^N\). For these manifolds Gompf introduced an invariant called \textit{genus-rank function} measuring the genera of smoothly embedded homologically essential surfaces in a given smooth 4-manifold. Also Gompf replaced the standard 2-handles with \textit{Casson handles} to define some exotic smooth structures on many handlebody interiors. Based on these results and techniques thanks to Gompf, the author of the present paper constructs exotic \(\mathbb{R}^4\)'s ``while maintaining control over their relationship with Stein surfaces''. The author constructs ``the first known'' examples of \textit{large exotic} \(\mathbb{R}^4\)'s that embed in Stein surfaces. Large \(\mathbb{R}^4\)'s are those that contain smooth compact codimension-0 submanifolds that do not embed in the standard \(\mathbb{R}^4\). The main theorems are complicated but their main consequences are summarized by the author in the introduction as follows. Theorem 1.1. There exist uncountably many large \(\mathbb{R}^4\)'s that each admit a smooth embedding into a Stein surface. Furthermore, these realize arbitrarily large (finite) values of the Taylor invariant by uncountably many compact equivalence classes. Theorem 1.2. There exist exotic \(\mathbb{R}^4\)'s that realize arbitrarily large (finite) values of the Taylor invariant and each contains a smooth compact codimension-0 submanifold that does not smoothly embed into any Stein surface. Theorem 1.3. If a smooth 4-manifold is obtained by end-summing some Stein surface with an \(\mathbb{R}^4\) defined by Theorem 1.1, then it still satisfies the adjunction inequality associated to this Stein surface. Theorem 1.4. Suppose that \(X\) is an open topological 4-manifold that is the interior of an oriented spin handlebody with all indices at most 2 and with \(0<\beta_2(X)<\infty\). There are smooth structures on \(X\) that realize infinitely many values of the Taylor invariant but all produce the same genus-rank function, and infinitely many genus-rank functions occur in this way. Similarly, there are smooth structures on \(X\) that produce infinitely many genus-rank functions but all realize the same arbitrarily large (finite) value of the Taylor invariant. Addendum 1.5. If the standard smooth structure that \(X\) inherits as a handlebody interior is compactly positive-definite, then each pair of Taylor invariant and genus-rank function from this theorem is realized by smooth structures on \(X\) representing uncountably many compact equivalence classes.