The engineering of materials with “properties-by-design” has spurred the creation of van der Waals (vdW) heterostructures that are based on the stacking of two-dimensional (2D) materials of varying compositions. These layered structures, held together by weak vdW forces, often show optoelectronic properties that are radically different from their individual building blocks. Typical synthesis of vdW heterostructures relies on large-scale chemical vapor deposition (CVD) or mechanical stacking of single 2D flakes.
Light drives the migration of charge carriers (electrons and holes) at the juncture between semiconductors with mismatched crystal lattices. These heterostructures hold promise for advancing optoelectronics and exploring new physics. The schematic’s background is a scanning transmission electron microscope image showing the bilayer in atomic-scale resolution. Credit: Oak Ridge National Laboratory, US Department of Energy. Image by Xufan Li/Chris Rouleau.
Recently reported in Science Advances (doi:10.1126/sciadv.1501882), Kai Xiao, a staff scientist at Oak Ridge National Laboratory (ORNL), and postdoctoral researcher Xufan Li, along with other co-workers from ORNL, Vanderbilt University, and Beijing Computational Science Research Center, presented the first known attempt to grow a misfit layer heterostructure containing GaSe and MoSe2 through a two-step CVD synthesis. The 2D heterostructures were fabricated by first reacting Se vapor with MoO3 to form monolayer MoSe2 crystals on SiO2/Si or fused quartz substrates. Once deposited, the as-synthesized MoSe2 (n-type) was then used to template the controlled growth of p-type GaSe to form a vertical misfit bilayer with no interfacial contamination.
Despite considerable success, “current CVD methods to directly grow 2D material heterostructures are limited to materials with similar lattice constants and/or crystal structures,” says Xiao of the Center for Nanophase Materials Sciences at ORNL. “It is a big challenge to put together two 2D materials with a large lattice-constant mismatch, but we are able to overcome this limitation with vdW epitaxial growth to create novel vdW heterostructures based on lattice-mismatched materials. This opens the door to new families of functional 2D materials for applications in photovoltaics, LEDs, transistors, and memory devices.”
The atomic structure of the bilayer heterostructures was characterized by scanning transmission electron microscopy, and the images exhibited repeating Moiré patterns, which hinted at long-range superlattice order. Investigation of the properties of the GaSe/MoSe2 misfit heterojunction showed efficient charge separation at the interface, which demonstrated strong coupling despite the 13% lattice mismatch, as revealed by photoluminescence quenching of the MoSe2. The p–n junction also showed a gate-tunable photovoltaic response, with the photoresponsivity ranging from 5.5 mA/W to 25.3 mA/W with increasing gate voltage.
“This vdW epitaxial growth method opens up a whole new family of heterostructures because you no longer need to worry about lattice matching at the unit cell,” comments Dave Johnson, an expert in misfit layer heterostructures at the University of Oregon. “The synthesis of heterostructures via vdW epitaxy, or other approaches, will change the way we think about materials optimization and enable the design of materials by actually modifying the identity, sequence, and nanoarchitecture of the heterostructured solid so that you can selectively tune device performance.”
This epitaxial growth protocol is not only limited to 2D semiconductors, but is predicted to also be accessible to both 2D metals and insulators. Despite this promise, the challenge will lie in tuning the growth temperature and pressure for each individual material in order to address problems with interfacial contamination, diffusion, and evaporation. If these conditions can be met, the synthesis of vdW misfit heterostructures will unlock a realm of materials that have not yet been explored, and will allow the selective tuning of materials’ properties based on its epitaxial building blocks.
