In theory, graphene should represent an ideal ultrathin barrier layer, as the pores between carbon atoms are smaller even than the radius of a helium atom. In practice, however, crystal boundaries and missing atoms allow vapour to permeate through the material, and the weak van der Waals bonds between planes mean that even stacks of multiple graphene layers can be penetrated.

The solution reported by Aria is to take a graphene monolayer formed by CVD, and to then use atomic layer deposition (ALD) to coat it with a 25–50 nm thick layer of alumina. Achieving conformal coatings on single-layer graphene is known to be difficult due to the material’s strong hydrophobicity. Aria found, however, that additional seed layers or pre-functionalizing of the graphene are not needed if the coating is applied immediately after the CVD stage, while the graphene–substrate complex is still hydrophilic, or if prolonged residence times are used to achieve optimal saturation conditions. The resulting nanoscale composite is suitable for metal passivation, device encapsulation and transparent barrier films.

Although a single layer of graphene paired with 50 nm of alumina does not achieve the extreme impermeability required for high-sensitivity applications like OLED encapsulation, the CVD–ALD process can be repeated until the necessary water vapour transmission rate (WVTR) has been reached. Barrier layers fabricated using this technique could exhibit appropriately low transmission rates with thicknesses of just tens of nanometres, compared to the millimetre-or-thicker layers used currently in televisions and smartphones.

Beyond its utility as a barrier material, graphene is of course in demand for microelectronics due to its optolectronic properties. Here, too, ALD alumina can be applied to coat and separate layers of active graphene in multilevel stacked devices. This means that delicate graphene structures can be protected during fabrication and processing, and their properties kept stable over time.

Impossible epitaxy

Petrov also reported the deposition of a metal-oxide layer on graphene, but in this case the material was strontium titanate (SrTiO3), with a view to fabricating a tuneable capacitor. The research reveals for the first time the growth mechanism of epitaxial oxide thin films on graphene transferred onto SrTiO3 and MgO substrates.

Petrov described how, after transferring a CVD graphene layer onto a SrTiO3 substrate, an additional 50 nm thick film of SrTiO3 was grown on top using reflection high-energy electron diffraction (RHEED) assisted pulsed laser deposition. High-resolution transmission electron microscopy (HRTEM) and x-ray diffraction (XRD) allowed the researchers to determine that the SrTiO3 nanolayer atop the graphene maintained an epitaxial relationship with the underlying substrate. Single-layer graphene prevents electronic interaction between the oxide layer and the substrate, so epitaxial growth of the upper nanolayer should have been impossible.

The answer, Petrov explained, lay in the initial local defects (such as grain boundaries) in the graphene layer, and the nature of the van der Waals bonds holding the graphene on the SrTiO3 substrate. The defects act like bridge-pillar spots, which enable the epitaxial growth of SrTiO3 across the graphene. This growth of SrTiO3 pillars also increases the interfacial sheer stress, resulting in partial folding of the graphene layer. Electrical testing of the fabricated capacitor structures showed that, despite the holes and multilayer patches, the graphene layer’s electrical properties were not compromised.