Publications

Semiconductor technology has enabled many essential devices, including transistors for computing and communication, diodes for solid-state lighting and photovoltaics. Behind all these devices, the material foundation is a series of highly elaborated heterostructures and superlattices [1]. To this end, integrating different materials into heterostructures or superlattices with well-defined spatial modulation of chemical compositions and electronic structures is central for generating designed electronic functions and has been a continued pursuit of the materials science community. The traditional approaches to heterostructures and superlattices, such as epitaxial growth, usually rely on the atom-to-atom covalent bonds to join the constituent materials and are often limited by strict lattice matching or processing compatibility requirements. High-quality heterostructures or superlattices can only be obtained between materials with nearly identical lattice structures and thus very similar electronic properties (e.g. Ga1−xAlxAs with slightly different compositions) [2]. A slight mismatch would inevitably lead to interfacial defects/strains, which could often propagate well beyond the interface to form extensive dislocations in bulk lattices, and in some cases, totally disordered interfacial layers (Fig. 1a and b) [3]. As a result, materials with substantially different lattice structures can hardly be integrated together without generating too many defects that may fundamentally alter their electronic functions.