Table of Contents
Fetching ...

Cohomologous symplectic forms with different Gromov widths

Shengzhen Ning

TL;DR

The paper resolves McDuff-Salamon's Problem 46 in dimensions $\ge 6$ by constructing closed manifolds with cohomologous symplectic forms that have different Gromov widths, and, in some cases, different first Chern classes. The strategy combines Li-Liu's symplectic cone theory for $b_2^+=1$, Biran's 4D ball-packings, and uniruledness to bound and distinguish widths, then uses diffeomorphism classifications (Wall, Jupp) to transport symplectic data to product manifolds $X\times S^2$ and $X'\times S^2$. The core construction starts with a non-diffeomorphic pair of homeomorphic 4-manifolds $X,X'$ and builds two cohomologous forms on $X\times S^2$ whose widths differ, with a further argument (Liu-Ohta-Ono) showing the first Chern classes must differ in this setting. Consequences include explicit examples such as $(\mathbb{C}P^2\#k\overline{\mathbb{C}P^2})\times S^2$ for $k\ge2$, and the method generalizes to higher dimensions via iterative products, providing negative answers to both problems in the stated regime.

Abstract

We study McDuff-Salamon's Problem 46 by showing that there exist closed manifolds of dimension $\geq 6$ admitting cohomologous symplectic forms with different Gromov widths. The examples are motivated by Ruan's early example of deformation inequivalent symplectic forms in dimension $6$ distinguished by Gromov-Witten invariants. To find cohomologous symplectic forms and compare their Gromov width, we make use of Li-Liu's theorem of symplectic cone for manifolds with $b_2^+=1$ and Biran's ball packing theorem in dimension $4$. Along the way, we also show that these cohomologous symplectic forms can have distinct first Chern classes, which answers another question by Salamon.

Cohomologous symplectic forms with different Gromov widths

TL;DR

The paper resolves McDuff-Salamon's Problem 46 in dimensions by constructing closed manifolds with cohomologous symplectic forms that have different Gromov widths, and, in some cases, different first Chern classes. The strategy combines Li-Liu's symplectic cone theory for , Biran's 4D ball-packings, and uniruledness to bound and distinguish widths, then uses diffeomorphism classifications (Wall, Jupp) to transport symplectic data to product manifolds and . The core construction starts with a non-diffeomorphic pair of homeomorphic 4-manifolds and builds two cohomologous forms on whose widths differ, with a further argument (Liu-Ohta-Ono) showing the first Chern classes must differ in this setting. Consequences include explicit examples such as for , and the method generalizes to higher dimensions via iterative products, providing negative answers to both problems in the stated regime.

Abstract

We study McDuff-Salamon's Problem 46 by showing that there exist closed manifolds of dimension admitting cohomologous symplectic forms with different Gromov widths. The examples are motivated by Ruan's early example of deformation inequivalent symplectic forms in dimension distinguished by Gromov-Witten invariants. To find cohomologous symplectic forms and compare their Gromov width, we make use of Li-Liu's theorem of symplectic cone for manifolds with and Biran's ball packing theorem in dimension . Along the way, we also show that these cohomologous symplectic forms can have distinct first Chern classes, which answers another question by Salamon.
Paper Structure (8 sections, 11 theorems, 30 equations)

This paper contains 8 sections, 11 theorems, 30 equations.

Key Result

Theorem 1.3

Let $X$ be any symplectic $4$-manifold homeomorphic but not diffeomorphic to the rational manifold $\mathbb{C}\mathbb{P}^2\#k\overline{\mathbb{C}\mathbb{P}^2}$ with $k\geq 1$. Then there exists some symplectic form $\omega_X$ on $X$ and $\omega_{S^2}$ on $S^2$ such that the $6$-manifold $M:=X\times

Theorems & Definitions (18)

  • Theorem 1.3
  • Corollary 1.4
  • Theorem 2.1: McDuffrationalruledMcduffimmersed
  • Theorem 2.2: LiLiu01
  • Theorem 2.3: biranpacking
  • Corollary 2.4
  • Remark 2.5
  • Remark 2.6
  • Definition 2.7
  • Theorem 2.8: Gromov
  • ...and 8 more