Constant Mean Curvature Surfaces in Homology Classes

Barış CoşkunuzerarXiv 2020 · DOI · arXiv

Abstract This paper proves that in any closed 3-dimensional manifold, non-trivial homology classes contain a continuous spectrum of constant mean curvature (CMC) surfaces. We demonstrate that there is no "gap" in the possible values of mean curvature, showing that if a class supports a surface with curvature $H_0$, it also supports surfaces for every value between 0 and $H_0$.

Introduction

In differential geometry, minimal surfaces (where mean curvature $H=0$) are classical objects representing area-minimizing configurations. A natural generalization is the Constant Mean Curvature (CMC) surface, which models physical systems like soap bubbles where internal pressure creates uniform curvature ($H > 0$). This research investigates the existence of CMC surfaces within specific homology classes of a 3-manifold, determining the range of curvature values $H$ for which such surfaces can be found.

Variational Methods and Metric Completion

The existence of CMC surfaces is established using a functional minimization approach. The author employs the functional $\mathcal{A}_H = |\Sigma| + 2H||\Omega||$, which balances the surface area $|\Sigma|$ with the volume $||\Omega||$ it encloses. Minimizers of this functional are surfaces with constant mean curvature $H$.

To facilitate the analysis, the manifold $M$ is "cut" along an area-minimizing minimal surface $\Sigma_0$. This process, known as metric completion, results in a new manifold $\widehat{M}$ with boundaries $\Sigma_0^+$ and $\Sigma_0^-$. This construction allows for a precise study of how surfaces deform as the curvature parameter $H$ varies.

The process of cutting a manifold $M$ along a minimal surface $\Sigma_0$ to create the manifold $\widehat{M}$ with boundaries.

Key Insights: The "No Gap" Lemma and Monotonicity

A central finding of this work is the "No Gap" Lemma, which states that the CMC spectrum of a homology class is a continuous interval $[0, \hat{H}]$. If a CMC surface exists for a curvature $H_0$, then CMC surfaces are guaranteed to exist for all $H \in [0, H_0]$.

Furthermore, the author proves a monotonicity property for these surfaces: as the mean curvature $H$ increases, the regions enclosed by the minimizing surfaces are nested ($\Omega_{H_1} \subset \Omega_{H_2}$ for $H_1 < H_2$). This nesting ensures that the surfaces move predictably through the manifold as "pressure" increases.

Illustration of nested regions $\Omega$ and $\Omega'$ and their corresponding surfaces $\Sigma$ and $\widehat{\Sigma}$ within the manifold $\widehat{M}$.

Main Results

The primary result (Theorem 1.3) asserts that for any closed embedded surface $S$ with nontrivial homology in a closed Riemannian 3-manifold $M$, there exists a range $[0, \hat{H}_{[S]}]$ such that for every $H$ in that range, there is a smoothly embedded CMC surface $\Sigma_H$ in the homology class of $S$.

The paper provides a concrete lower bound $C_{[S]}$ for this threshold, calculated using the "girth" of the homology class—the maximum area required to sweep through the class—and the area of the initial minimal surface. Additionally, the author shows that for generic values of $H$ within the interval, the minimizing CMC surface is unique.

Conclusion and Future Directions

While this work establishes a robust existence theory for stable CMC surfaces, the provided lower bounds are likely not sharp. The author suggests that min-max methods, which identify unstable "saddle point" surfaces, could potentially extend the known spectrum of CMC surfaces. Future work will focus on proving Conjecture 4.6, which proposes a stronger lower bound based on the "width" of the manifold.

Summary updated

Abstract

We show the existence of constant mean curvature surfaces in the homology classes of closed 3-manifolds.
BibTeX
@misc{coskunuzer2020constant,
  author = {Barış Coşkunuzer},
  title = {{Constant Mean Curvature Surfaces in Homology Classes}},
  year = {2020},
  eprint = {2001.00505},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2001.00505},
}
Figure