Ferroelectric materials have a spontaneous electric polarization that can be switched by an applied field. Traditionally, ferroelectricity requires a crystal structure with broken inversion symmetry — specific atomic arrangements where positive and negative charges don't balance locally. The chemistry dictates the electric order.
In twisted square bilayers, the twist alone is sufficient.
Take two layers of a centrosymmetric material — one that has no ferroelectricity whatsoever in its bulk form — and rotate them by a small angle. The moiré pattern created by the misalignment breaks inversion symmetry locally. Within each moiré unit cell, the stacking arrangement varies smoothly from one registry to another, and some of these registries have a net out-of-plane dipole moment. The result: a ferroelectric polarization that exists because of geometry, not chemistry.
Note (2026-11-14; read against the paper's full text): the essay says the twist alone breaks inversion symmetry in a centrosymmetric material, that no atoms move on switching, and that switching needs modest fields because the moire domains are soft. The paper says the moire ferroelectricity originates from sliding ferroelectricity in the untwisted aligned bilayers, where stacking-dependent interlayer charge transfer gives an out-of-plane polarization, for the tetragonal layer groups P-42m and P-4m2, in which its symmetry analysis allows that polarization; the words centrosymmetric and inversion do not appear in it. The switching it describes is between a single-band and a layer-resolved bilayer Hubbard model, set by the competition between the ferroelectric layer-asymmetric potential and interlayer tunneling, and its representative materials are Cu2WS4 (ferroelectricity-dominated) and GeCl2 (tunneling-dominated); I found nothing about domain softness or switching fields.
The polarization then reaches back into the electronic structure. The electric field from the ferroelectric pattern shifts the on-site energies differently at different points in the moiré cell, reshaping the electronic bands. Flat bands — which host correlated electronic phases — can be tuned by switching the ferroelectric state. The mechanical arrangement produces the electric order, and the electric order produces the band engineering.
The switchability is the remarkable part. In a conventional ferroelectric, switching requires moving atoms — overcoming energy barriers set by the crystal structure. In the moiré ferroelectric, switching requires changing the interlayer charge distribution, which can be done with modest electric fields because the moiré stacking domains are soft. The switching energy is set by the moiré, not by atomic bonds.
No atoms move. The twist is fixed. What switches is the charge distribution within the moiré cell — a purely electronic reconfiguration within a geometrically imposed landscape.