Category:2D materials: Difference between revisions

Every real crystal or material has a surface. In VASP simulations, however, periodic boundary conditions typically model an infinite crystal, the perfect bulk. To model a surface, a thin film, or an intrinsically 2D material in VASP, therefore, requires breaking these periodic boundary conditions intentionally. This is done by elongating the simulation cell in one direction (normal to the intended surface) without adding more atoms. This creates a vacuum region between repeated images of thin films of material (commonly called surface slabs, or just slabs), each of which has two surfaces.

Nomenclature

Surfaces

As discussed above, in atomistic simulations employing periodic boundary conditions, surfaces are modeled as thin films of a material separated by a vacuum region of sufficient thickness. When the surface itself is the primary focus, the slab model must be constructed to mimic a semi-infinite bulk crystal beneath the surface. The second surface, located on the opposite side of the slab, should be sufficiently separated to prevent strong interactions between the two surfaces.

Thin films

Although thin-film simulations often employ the same computational setup and slab model as surface calculations, their objective differs. In thin-film studies, the goal is to investigate the behavior and properties of the entire system — including both surfaces and the bulk-like interior region — rather than isolating the characteristics of a single surface.

2D materials

This class of materials comprises ultrathin films consisting of only a few atomic layers. Prototypical examples include graphene, which is one atomic layer thick, and molybdenum disulfide (MoS$_2$), which consists of three atomic layers. Such materials can occur naturally as van der Waals–bonded layered crystals, but they can also be exfoliated and studied as mono-, bi-, or few-layer systems.

Creating slab models

A clean surface is defined uniquely by the bulk crystal, the lattice plane where the cut should be placed, and the termination of the surface. For a slab model, additional parameters are important, mainly the slab thickness and the thickness of the vacuum region. Both need to be converged independently of each other to be sure to get accurate results.

While it is entirely possible to create a slab model by visualizing a crystal and thinking hard about lattice planes and stacking, it is usually more convenient and less error-prone to use available tools. The ASE^[1] and pymatgen^[2] both provide surface building capabilities. The following ASE code snippet creates a NaCl 001 slab with 6 layers for example:

from ase.build import bulk, surface
from ase.io import write
nacl_bulk = bulk('NaCl', 'rocksalt', a=5.64, cubic=True)
slab_001 = surface(nacl_bulk, (0,0,1), 3)
slab_001.center(vacuum=10.0, axis=2)
write('POSCAR_NaCl_001_6layers.vasp', slab_001, format='vasp', direct=True, sort=True)

Related pages

• Computing the work function
• Electrostatic corrections
• IDIPOL
• CO on Ni 111 surface