Subdividing three-dimensional Riemannian disks
Papasoglu asked whether for any Riemannian 3-disk [Formula: see text] with diameter [Formula: see text], boundary area [Formula: see text] and volume [Formula: see text], there exists a homotopy [Formula: see text] contracting the boundary to a point so that the area of [Formula: see text] is bounded by [Formula: see text] for some function [Formula: see text]. He further asks whether it is possible to subdivide [Formula: see text] by a disk [Formula: see text] into two regions of volume [Formula: see text] so that the area of [Formula: see text] is bounded by some function [Formula: see text]. In this paper, we answer the questions above in the negative: We prove that given [Formula: see text] and [Formula: see text], one can construct a metric [Formula: see text] so that any 2-disk [Formula: see text] subdividing [Formula: see text] into two regions of volume at least [Formula: see text], the area of [Formula: see text] is greater than [Formula: see text]. We further prove that for any Riemannian 3-sphere [Formula: see text], there is a surface that subdivides the disk into two regions of volume no less than [Formula: see text], and the area of this surface is bounded by [Formula: see text], where [Formula: see text] is the homological filling function of [Formula: see text].