Answers are grounded in this paper and cite their sources.
For a fixed n ≥ 6, no closed polyhedral surface in three-dimensional space, of any genus, orientable or not, embedded or immersed, has only planar convex regular n-gons as faces. For n = 3, 4, 5 such surfaces exist in infinitely many shapes. The proof rests on a sharp statement about the sphere: for 2π/3 ≤ α < π no simple closed spherical polygon with all sides of length α lies in an open hemisphere, whereas for 0 < α < 2π/3 the equilateral triangle does. For sides of two lengths this fails: for all 2π/3 ≤ b < a < π some simple closed spherical quadrilateral with sides a, a, b, b lies in an open hemisphere, so, for instance, two regular heptagons and two regular hexagons can share a vertex that is their unique lowest point. In general, if all sides of a simple closed spherical polygon in an open hemisphere are at least 2π/3 long, then for every generic pole one of them lies directly above a longer side over the interior of its longitude range. Whether regular faces of several kinds, all with six or more sides, can form a closed polyhedral surface remains open. MSC 2020: Primary 52B70; Secondary 52B10, 51M20, 52C25. Files: the paper (PDF), its LaTeX source, a reproduction archive with the program that generates the figures and a script that re-evaluates the examples numerically, and a Lean 4 archive with formal proofs of three statements of the paper (Section 8 of the paper says which ones).
No takes yet. Share an insight, caveat, or question.
Sungsoo Na (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: