The frictional behavior of encapsulated graphene nanoribbons is pivotal for enabling ultralong growth and ensuring the reliability of nanoelectronic devices, yet it remains elusive due to the buried nature of the interfaces. Using large-scale molecular dynamics simulations, we demonstrate that GNRs confined between hexagonal boron nitride (h-BN) layers exhibit friction forces over an order of magnitude lower than those of their on-surface counterparts. The kinetic friction force ( $F_{\mathrm{k}}$) scales sublinearly with length ( $L_{\text{GNR}}$) as $F_{\mathrm{k}}\propto {(\ln \left(L_{\text{GNR}}\right))}^3$. This robust superlubric state originates from interfacial registry competition with adjacent AA′-stacked h-BN layers, which inhibits coherent strain accumulation and suppresses stick-slip behavior, thereby promoting collective low-dissipation sliding. Furthermore, we reveal a negative differential friction coefficient versus confinement pressure as the number of h-BN layers increases, arising from the competition between pressure-induced ripple suppression and the monotonically enhanced h-BN/h-BN interlayer perturbations in thicker stacks. This interplay provides a practical avenue for tuning friction via confinement engineering. These findings elucidate atomistic dissipation pathways at buried van der Waals interfaces and establish encapsulation as a viable route to robust GNR-based nanoelectronics and scalable superlubric systems.

One-dimensional (1D) nanomaterials easily bend due to perturbations from their surroundings or their own behaviors. This phenomenon not only impacts the performances of various devices but has also been employed to develop a variety of new functional devices, in which the bending energies of the nanomaterials determine the device performances. However, measuring the energies of such nanomaterials is extremely difficult. Herein, pseudo-break imaging of 1D nanomaterials has been proposed and realized on individual carbon nanotubes (CNTs), in which a CNT appears to break and has a fracture but is actually intact. This imaging approach provides the values of the bending energies of the CNTs with an accuracy of 1–50 eV. Furthermore, this imaging approach can manipulate the bending shapes and energies of CNTs. This work presents a protocol for bending analysis and manipulation, which are vital to fundamental and applied studies of 1D nanomaterials.