Answers are grounded in this paper and cite their sources.
Twisted van der Waals structures exhibit moiré superlattices in which lattice mechanics are intrinsically coupled to electronic structure. At marginal twist angles, elastic relaxation produces periodic networks of stacking domains separated by nanometer-wide soliton boundaries. In multilayer graphene, these domains may correspond to Bernal (ABA) and rhombohedral (ABC) stacking, which exhibit distinct electronic and correlated behaviors [1]. The solitons act as topological domain walls characterized by localized interlayer displacement that accommodates twist mismatch. Although often treated as pure shear boundaries, they can also contain tensile and compressive components depending on local constraints and structural heterogeneity [2]. Functioning as quasi–one-dimensional electronic channels, these soliton boundaries strongly modulate band structure, correlated phenomena, and emergent polar or topological textures. The ability to dynamically reconfigure such networks through mechanical perturbation therefore offers a route toward programmable quantum phases and strain-tunable quantum functionality. Despite extensive characterization of static moiré structures, dynamic mechanical control over soliton networks remains largely unexplored. Prior studies have investigated the atomic-level reconstruction, and scanning probe techniques have demonstrated that both electric field and local out-of-plane forces can create or reposition individual soliton boundaries [3,4,5]. However, such approaches are inherently localized and quasi-static, offering limited temporal resolution and restricted access to collective, strain-driven dynamics. Real-time observation of in-plane strain-induced domain-wall motion has therefore remained elusive. Controlled in situ straining within the transmission electron microscope (TEM) enables direct interrogation of strain-driven motion, providing mechanistic insight into soliton dynamics and informing strategies for the rational design of in-plane strain-programmable devices. Here, we demonstrate mechanical reconfiguration of large-scale moiré soliton networks in twisted double bilayer and trilayer graphene (tDBLG/tDTLG) through MEMS-based in situ straining in TEM. Samples were first fabricated via the tear-and-stack method to produce controlled twist angles (∼0.1°), and then integrated onto push-to-pull (PtP) devices using a customized transfer strategy to preserve structural integrity (Fig. 1a). The PtP architecture converts actuator displacement into uniaxial tensile or shear strain across the gap region, enabling controlled in-plane loading of the suspended materials. Twist angle and stacking domains were verified prior to loading by selected-area diffraction (SAED) and dark-field TEM (DF-TEM). By selecting first- or second-order graphene diffraction spots, soliton boundaries were isolated from the hexagonal boron nitride (hBN) background and imaged with high sensitivity, with simulated and experimental contrast shown in Fig. 1b–e. An overview dark-field image of the device-mounted suspended structure is shown in Fig. 1f. Application of tensile and shear strain in TEM enables the first direct real-time observation of collective soliton dynamics (Fig. 2a–c shows one representative video), revealing soliton-wall motion, junction migration, and large-scale network reorganization, all of which include rapid switching between Bernal and rhombohedral stacking configurations. Machine-learning-assisted tracking of AA stacking junctions enables robust extraction of junction trajectories, from which geometric analysis quantifies anisotropic soliton motion. In parallel, strain inferred from edge tracking can be directly compared with the applied load recorded by the holder, revealing elastic versus irreversible regimes throughout loading cycles. Together, these measurements establish a quantitative framework for correlating applied strain with soliton behavior in two-dimensional (2D) quantum materials. Beyond the present system, the in situ straining architecture developed here provides a broadly applicable platform for mechanically probing and reconfiguring a wide range of stacked 2D materials. Future integration of Lorentz TEM, high-resolution scanning transmission electron microscopy (STEM), four-dimensional scanning transmission electron microscopy (4D-STEM), and electron energy loss spectroscopy (EELS) will enable direct correlation of externally applied mechanical stress with magnetic textures, atomic-scale configurations, local strain fields, and electronic structure evolution. Through such multimodal approaches, we aim to provide deeper insight into strain-mediated topology and electronic reconstruction, advancing strategies for deterministic control of moiré networks in programmable quantum materials [6]. Measurement geometry and representative dark-field TEM images. (a) Schematic illustration of the MEMS-based in situ straining setup inside the TEM. The black arrow denotes the direction of applied uniaxial tensile strain. (b–c) Experimental and simulated dark-field images from a representative first-order Bragg reflection. (d–e) Experimental and simulated dark-field images from a representative second-order Bragg reflection. (f) Cropped frame from an in situ uniaxial tensile straining video. Scale bar = 500 nm. The frame is Fourier filtered to enhance soliton contrast. White arrows indicate the strain direction. Collective soliton motions under in situ shear loading. (a–c) Sequential frames from an experimental dark-field TEM video recorded during in situ shear loading, highlighting collective soliton dynamics. Scale bar = 100 nm. White arrows indicate the strain direction.
No takes yet. Share an insight, caveat, or question.
Wang et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: