On the boundaries of highly connected, almost closed manifolds

TL;DR

This paper proves new results on the boundaries of highly connected, almost closed manifolds, settling longstanding questions, classifying Stein fillable spheres, and analyzing Toda brackets through advanced spectral sequence techniques.

Contribution

It establishes that certain boundaries of highly connected, almost closed manifolds also bound parallelizable manifolds, resolving key conjectures and advancing manifold classification.

Findings

Boundaries of certain manifolds also bound parallelizable manifolds.

Resolved longstanding questions of C.T.C. Wall.

Proved a conjecture of Galatius and Randal-Williams.

Abstract

Building on work of Stolz, we prove for integers $0 \le d \le 3$ and $k>232$ that the boundaries of $(k-1)$-connected, almost closed $(2k+d)$-manifolds also bound parallelizable manifolds. Away from finitely many dimensions, this settles longstanding questions of C.T.C. Wall, determines all Stein fillable homotopy spheres, and proves a conjecture of Galatius and Randal-Williams. Implications are drawn for both the classification of highly connected manifolds and, via work of Kreck and Krannich, the calculation of their mapping class groups. Our technique is to recast the Galatius and Randal-Williams conjecture in terms of the vanishing of a certain Toda bracket, and then to analyze this Toda bracket by bounding its $\mathrm{H}\mathbb{F}_p$-Adams filtrations for all primes $p$. We additionally prove new vanishing lines in the $\mathrm{H}\mathbb{F}_p$-Adams spectral sequences of spheres and Moore spectra, which are likely to be of independent interest. Several of these vanishing lines rely on an Appendix by Robert Burklund, which answers a question of Mathew about vanishing curves in $\mathrm{BP} \langle n \rangle$-based Adams spectral sequences.