On the curvature bounded sphere problem in $\mathbb{R}^3$
Hongda Qiu
TL;DR
The paper investigates whether a topological sphere $S$ smoothly embedded in $\mathbb{R}^3$ with normal curvatures bounded by $1$, contained in a ball of radius $2$, must enclose a region containing the unit ball of radius $1$; this addresses a Burago–Petrunin conjecture about minimal enclosed volume $\frac{4}{3}\pi$. It proves a stronger radius-$2$ result by showing the enclosed region is star-shaped via a short radial projection from the bounding sphere and leveraging a curvature-based distance estimate to guarantee a contained unit ball. The work clarifies the landscape of the Burago–Petrunin problem by isolating a finite-radius case, discusses the tightness of the radius bound, and presents a genus-$2$ counterexample construction in the appendix to illustrate potential obstacles. Overall, the results contribute a concrete geometric criterion linking normal curvature bounds and ambient containment to the presence of a unit ball inside the enclosed region, with implications for potential positive or negative resolutions of the conjecture.
Abstract
We prove that if a topological sphere smoothly embedded into $\mathbb{R}^3$ with normal curvatures absolutely bounded by $1$ is contained in an open ball of radius $2$, then the region it bounds must contain a unit ball. This result suggests a potential direction for a problem formulated by D.Burago and A.Petrunin asking whether a topological sphere smoothly embedded in $\mathbb{R}^3$ with normal curvatures absolutely bounded by $1$ encloses a volume of at least $\frac{4}{3}π$. The appendix presents an example illustrating an alternative aspect for this problem.
