Weak van der Waals interactions drive 2D material properties
Germany-based researchers discuss how weak van der Waals interactions have a strong impact on the properties of 2D materials.
Two-dimensional (2D) crystals are highly ordered covalently or ionically bonded structures with a sheet thickness of only one or a few atoms. Their parental materials are layered structures with the individual 2D crystalline layers coupled by weak van der Waals forces. Crystalline structures in monolayers of biomolecules have been known for years, but truly 2D solid-state crystals have been expected to be unstable when isolated from bulk crystals.
Only 15 years ago, the cleavage of graphite into single atomic layers, called graphene, started a new research area on monolayer and few-layer crystals of 2D materials, also called van der Waals materials. After the discovery of graphene – the most prominent representative of this class of materials – many other 2D crystals have been identified, often with intriguing properties that have no counterparts in 3D solids.
2D materials cover a wealth of physical, mechanical and chemical properties. In addition to the materials classes known from 3D, such as insulators, semiconductors and metals, 2D materials can also be ferromagnetic, ferroelectric or superconducting, and can host topological and correlated electronic phases. They can be single photon sources, being super strong and flexible, inert and (photo-)catalytically active.
The real advantage of these materials is they can be arbitrarily combined by mechanical stacking without the constraints of the out-of-plane registry, as with conventional 3D solids. Due to proximity and hybridisation effects, the assembly in such new artificial van der Waals solids can result in the emergence of new states of matter with novel functionalities that not only depend on the combination of different 2D crystal but also on their rotational alignment.
One compound, two materials
In the early days of 2D materials research it was discovered that many 2D crystals have significantly different properties to their parent bulk phase. For example, the Dirac cone characterising graphene is absent in graphite. Group 6 transition metal dichalcogenides (TMDCs), the most researched class of transition metal chalcogenides which includes molybdenum disulphide (MoS2) and tungsten diselenide (WSe2), shows significant band gap opening when going from bulk to monolayer, and an indirect-to-direct band gap transition from bi to monolayer.
Also, symmetry effects have been studied. Most Group 6 TMDCs show 2H symmetry – two layers per hexagonal unit cell – in the bulk with inversion symmetry centres in between the layers. Thinned down to the monolayer, their inversion symmetry gets broken and a significant spin-orbit gap opens in the structure – the K points in the Brillouin zone. This can be exploited by what is called valleytronics, where the spin-polarised electrons in the different valleys can be accessed independently using spin-polarised light. In the bilayer, this strong spin polarisation vanishes, as the inversion symmetry is re-introduced, so bilayer and monolayer have very different spin physics.
Opening the band gap is a general phenomenon when thinning down a layered material to a 2D crystal, and can be rationalised by the stronger quantum confinement of the electrons in the former. So what happens to 2D crystals made from metallic parent layered bulk materials, like noble metal chalcogenides?
The most researched noble metal dichalcogenide is platinum diselenide (PtSe2), which has been investigated as a layered material since the late 1950s, proposed as an interesting 2D crystal in 2014, and manufactured one year later by the selenisation of a metallic platinum – Pt(111) – surface. The metallic bulk becomes semiconducting when it is thinned down to three layers, and further thinning opens an appreciable band gap. It has been shown that it is very sensitive to molecular adsorption and shows remarkable piezoelectric and magnetoresistive effects.
Such metal-semiconductor transitions can be interesting for developing single-compound devices, for example, the transitional setup of a 2D transistor where a metallic contact interfaces the 2D crystal, could be replaced by a single or multilayer system of the same material. This would lead to superior energy efficiency due to the reduction of the Schottky barrier between metal and semiconductor, and facilitate recycling of the device.
Magnetic ordering in 2D is prohibited by the Mermin-Wagner theorem, but correlation between electron spins can be mediated by spin-orbit coupling. Ferromagnetic ordering was demonstrated experimentally in 2017 with chromium triiodide (CrI₃), of which the parent 3D bulk crystal is also ferromagnetic, maintaining this property for the monolayer. However, in the bilayer the coupling becomes antiferromagnetic, while ferromagnetism is recovered in the triple layer, indicating intriguing layer-dependent magnetic properties.
Other research has demonstrated 2D materials turning into superconductors when cooled down to cryogenic temperatures, for instance nickel diselenide (NiSe₂) with a layer-dependent superconducting transition temperature with the superconductivity present even in the monolayer limit.
Even more fascinating, some semiconducting Group 6 TMDCs, including MoS₂, exhibit a superconducting phase when highly charge doped. Doping can be realised in a field-effect transistor structure known from standard electronic applications facilitating in-situ manipulation of phase transition. This electric field tunable superconductivity has the potential for the design of complex circuitries with optically, electronically active and spin-selective regions as well as superconducting parts that can provide single phonon emission, manipulation and detection in one monolithic device.
Many 2D materials have been synthesised and many more experimentally explored. The most common approach is the synthesis of bulk van der Waals crystal followed by mechanical exfoliation in single and few-layered crystals, with the subsequent assembly into complex van der Waals stacks by dedicated micromechanical manipulation with sub-micrometer lateral precision, rotational alignment better than 0.1° and ultra-clean interfaces.
Apart from the manifold options for different material combinations, stacking sequences and tailored interfaces, another degree of freedom unique to van der Waals structures is that the rotational angle between individual layers has a strong impact on the properties of 2D heterostructures, and thus opens the door to a new research field called Twistronics.
The rotational degree of freedom and small differences in the lattice constant of the stacked 2D lattices causes in-plane moiré superlattices. Typically, they have the same lattice as the individual crystals, but with a larger lattice constant that depends on the twist angle between the two layers and the difference in lattice constant.
An interesting phenomenon is the stacking of two honeycomb lattices with an angle of exactly 30° creates a quasicrystalline superlattice. Moiré physics offers the possibilities for engineering electronic bandstructure and optical properties, but also results in emergent phases of matter due to altered electronic, spin and valley interactions. Striking examples are that two layers of the semi-metal graphene can be turned into a superconductor by stacking them under a magic angle of 1.1°, or the formation of manifold interlayer excitons by stacking two 2D semiconductors under near 0° or near 60°.
Artificial van der Waals solids
An interlayer exciton is an electron-hole pair that is bound by Coulomb forces, with the electron located in one layer and the hole in the other layers. Our research found this image describes the underlying physics very well for charger carriers with a certain energy and momentum relationship, but it fails for electron-hole pairs with a different energy and momentum relationship, indicating that these electrons and holes are located at different valleys in momentum-space.
It appears that even if the van der Waals coupled layers have saturated atomic orbitals on the surface, hybridisation between the layers occurs for certain material combinations with special rotational alignment and sufficiently clean interfaces for certain regions in momentum-space. This valley selective hybridisation results in the formation of an ‘artificial’ van der Waals solid with modified electronic and optical properties. Charge carriers residing at other valleys show the characteristic of a conventional semiconductor heterojunction – the separation of the electron-hole pair as essential for efficient photovoltaic devices.
Further artificial van der Waals solids can be created by making use of proximity effects. These include those between magnets, superconductors, semiconductors or materials with non-trivial topologies, where the emergence of interaction driven new phases of matter and novel presumably exotic quantum states are expected.
In chemistry, 2D polymers – a completely new class of 2D crystals – has emerged. We are currently investigating how they impact a 2D crystal if interfaced in a heterostructure.
Challenges and perspectives
The theoretical description of these van der Waals (hetero)structures becomes increasingly complicated. Direct first-principles calculations, as for traditional 2D crystals, becomes impossible because the model structures become unfeasibly large even for the fastest supercomputers, as super cells can quickly contain some 10,000 atoms.
To make things even more complicated, the workhorse of ab intitio theory, density-functional theory (DFT) becomes unreliable for van der Waals interactions and for strong correlations, even though there are practical workarounds to reduce the impact of these issues. The 2D boundary conditions mark an additional technical difficulty. Methods beyond DFT, including more accurate but computationally very demanding methods, can accurately describe these interactions, but are more costly. Thus, approximate methods, model hamiltonians and analytical tools have to be employed to describe the physical effects in these materials, and thoroughly validated to allow predictions on new, interesting materials with intriguing properties.
To exploit the exciting properties of 2D materials and their heterostacks for realistic applications, the large-scale synthesis of high-quality crystals is essential, however, so far this is lacking. There are three routes for the synthesis of 2D materials:
I) van der Waals growth on flat inert substrates by metalorganic chemical vapour deposition, physical vapour deposition, pulsed laser deposition or molecular beam epitaxi approaches. The challenge here is either the transfer from the growth substrate to a target substrate without surface and interface contamination and mechanical ‘wrinkles’, or van der Waals growth itself.
In van der Waals growth, multiple nucleation sites on a large substrate typically cause polycrystalline materials with crystallites ranging from a few to a few hundreds of micrometers. The control of the number of layers during growth over a large range is another challenge. Nevertheless, this approach is
compatible with well-established technologies in semiconductor industries.
II) Liquid exfoliation using an intercalation approach or application of mechanical forces, such as by sonication with subsequent centrifugation to select crystallites of a certain thickness and size in solution. This ink containing 2D crystallites with a typical size of a few micrometers can then be used for ink printing spray or dip coating of a thin network of 2D crystals. Even if the properties of such networks do not match those of monocrystalline films, they can be good enough for applications in the area of electronics, optics or solar energy conversion. This process is scalable, cheap and paves the way for printable on-demand electronics on flexible substrates as well as the development of functional inks and paints.
III) Micromechanical exfoliation from bulk crystal using adhesive tapes and the subsequent transfer on a target substrate or the van der Waals assembly into heterostacks using, for example, a viscoelastic dry transfer process with help of mechanical translations and rotations-stages for precise alignment. The first successful synthesis of graphene monolayers has been realised using micromechanical exfoliation. While very time consuming, not scalable and the parameters are hard to control, it is still preferred for producing 2D crystals and heterostacks with superior quality in fundamental research. The approach enables almost unlimited flexibility and speed of production.
Progress is highly likely for the vibrant and young research field of 2D materials and van der Waals assemblies. At this stage, we can only speculate on the first marked application of van der Waals stacks, however, strong candidates for all 2D applications on the market are already present in the developing field of quantum technologies, in sensing and solar energy applications. For example, they could help realise efficient and entirely transparent solar cells, and combining several 2D materials for their sensitivity to environmental atoms molecules may make them useful for engineering flexible, transparent, robust, stable and high-sensitivity gas sensors.
To explore these open questions and challenges, the German Research Foundation (Deutsche Forschungsgemeinschaft) has initiated a priority programme and will fund up to 30 researchers to investigate the physics that emerges from weak interlayer interactions of 2D materials. Research will be supported for a six-year period starting in summer 2020.
Thomas Heine is Chair of Theoretical Chemistry, Technische Universität, Germany and Ursula Wurstbauer is Full Professor at the Institute of Physics, Westfälische Wilhelms-Universität, Germany.