Link to the paper https://www.nature.com/articles/s41563-018-0176-4

Our findings confirm that transparency of 2D materials highly depends on the binding nature of 2D materials as well as substrates. Such understanding enables the growth of various single-crystalline thin film materials on 2D material-coated substrates, and the release of these functional thin films as freestanding membranes provides the unique opportunities for heterointegration of arbitrary thin film single-crystalline materials.

Production and application of epitaxial semiconductor thin films has been an interesting challenge for the semiconductor industry for several decades. Silicon, despite its indirect bandgap, relatively low electron and hole mobility, and poor thermal characteristics - has been used extensively in electronics due to its low cost and easy producibility. Higher performance materials such as GaAs, GaN, and more are not as widely used due to the high production costs and associated processing challenges.

To tackle this issue, our group has developed the process of remote epitaxy, which allows producing single-crystalline layers of epitaxial III-V and III-N materials. In this fashion, we produce layers of the required thickness for device applications, and re-use the substrate for subsequent growth. The technique shows promise of significantly reducing the cost of high-performance semiconductors by minimizing the need for expensive bulk wafers.

Remote epitaxy leverages both the electronic properties and the surface properties of 2D materials. Layer(s) of 2D material are transferred onto a single-crystalline substrate (for example, GaAs, as demonstrated previously by Kim et. al.). Then, using epitaxial growth methods such as metal-organic chemical vapor deposition (MOCVD), an epitaxially-aligned crystalline layer is produced on top of the 2D material. Then, this layer can be exfoliated with a mechanical handler, as a result of the weak van der Waals interaction between the 2D material and substrate layer.

Our study aims to shed light on possible mechanistic routes by which the remote epitaxy process is possible, and establishes a set of rules to describe remote epitaxial interactions. We used density functional theory (DFT) for a systematic analysis of interactions, with corresponding experimental validation. Our study focused on a few variables, which include:

• Number of 2D material monolayers on the guiding substrate
• Ionicity of the material (i.e. degree of ionic character in the bonds in the material)
• Ionicity of the 2D material interlayer

What we found through DFT was that graphene (a 0% ionicity layer) imprints the potential energy landscape of the underlying substrate, and that this effect was enhanced with more strongly ionic substrates (such as LiF). However, the substrate energy landscape is screened as more 2D interlayers are added, which means that for three layers of graphene, only LiF (the strongest ionic compound) yielded single-crystalline layers. However, for a high-ionicity layer such as h-BN, the energy landscape interferes with the substrate landscape, resulting in poly-crystalline epitaxial layers. As the number of hBN layers is increased however, the epitaxial layer follows the energy landscape of h-BN rather than the substrate, resulting in single-crystalline layers – which contradicts the behavior of graphene.

In summary, the DFT calculations and following experiments revealed that the ionicity of the substrate and the 2D interlayer dictate the energy landscape available for epitaxial materials to grow on. This energy landscape determines the crystallinity of the epitaxial layer. These results will allow the promising route to realize low-cost high performance electronics and optoelectronics.