Switching Two-Dimensional Sliding Ferroelectrics by Mechanical Bending.

Two-dimensional van der Waals materials, possessing a unique stacking degree of freedom, offer an alternative strategy for modulating their properties through interlayer sliding. Controlling the stacking order is crucial for tuning material properties and developing slidetronics-based devices. Here, using machine-learning potentials, we propose a mechanical bending approach to manipulate stacking orders and related properties in sliding ferroelectric h-BN, 3R-MoS_{2}, and nonferroelectric bilayer graphene. Our simulations predict the formation of irreversible kinks in bent bilayers, deviating from the expected arclike deformation. This kink formation arises from the interplay between bending energy and interlayer stacking energy. Notably, the bending-induced kink contains a ferroelectric topological domain wall that reverses the polarization of sliding ferroelectrics, a mechanism distinct from the conventional flexoelectric effect. This work proposes an exciting mechanical bending approach to dynamically manipulate the stacking order and associated optical, topological, ferroelectric, and magnetic properties in van der Waals layered materials.