The accurate molecular dynamics simulation of weakly bound adhesive complexes, such as supported graphene, is challenging due to the lack of an adequate interface potential. Instead of the widely used Lennard-Jones potential for weak and long-range interactions we use a newly parameterized Tersoff-potential for graphene/Ru(0001) system. The new interfacial force field provides adequate moiré superstructures in accordance with scanning tunnelling microscopy images and with DFT results. In particular, the corrugation of ξ ≈ 1.0 ± 0.2Å is found which is somewhat smaller than found by DFT approaches (ξ ≈ 1.2Å) and is close to STM measurements (ξ ≈ 0.8 ± 0.3Å). The new potential could open the way towards large scale simulations of supported graphene with adequate moiré supercells in many fields of graphene research. Moreover, the new interface potential might provide a new strategy in general for getting accurate interaction potentials for weakly bound adhesion in large scale systems in which atomic dynamics is inaccessible yet by accurate DFT calculations. keywords: atomistic and nanoscale simulations, molecular dynamics simulations, corrugation of graphene, moire superstructures
INTRODUCTIONSince the reliable ab initio density functional theory (DFT) methods with various van der Waals (VDW) functionals can be used only up to ∼ 1000 atoms [1] the development of an adequate classical interfacial force-field for supported graphene (gr) is crucial. Although, classical molecular dynamics (CMD) simulations can not be used for the study of electronic structure, however, many important properties of graphene, such as surface topography, moire supercells and interfacial binding characteristics, adhesion as well as defects and many other features can in principle be studied by CMD simulations [2, 3] without the explicit consideration of the electron density or orbitals.The size-limit of DFT is especially serious if accurate geometry optimization or molecular dynamics simulations should be carried out. Moreover, various VDW DFT functionals provide diverging results [1, 4]. In particular, accurate DFT potential energy curves for the interface of gr-substrate systems has been obtained only by the extremely expensive random phase approximation (RPA) for small modell systems which fail, however, to include the full moire-superstructure [4][5][6]. Using CMD simulations no size-limit problem occurs. An important limitation here, however, is the limited availability of accurate interface potential for weak interactions. The widely accepted choice is the simple pairwise Lennard-Jones (LJ) potential. In the present paper we point out that this potential is inadequate, however, for gr/Ru(0001) due to the improper prediction of the binding sites (registry). Therefore, the development of reliable interatomic potentials for gr-support systems is inevitable.First principles calculations (such as density functional theory, DFT) have widely been used in the last few years to understand corrugation of nanoscale gr sheets on various substra...