Nanoindentation Response of Polycrystalline Graphene by Atomistic Simulations
Ying Lu,Jin Qian
DOI: https://doi.org/10.2991/ame-16.2016.36
2016-01-01
Abstract:Nanoindentation has been widely used to determine the mechanical properties of pristine and polycrystalline graphene in experiments. To investigate the effects of grain size on the mechanical properties of polycrystalline graphene, we use molecular dynamics simulations to study the nanoindentation response of suspended polycrystalline graphene layer. The force-displacement behavior from the simulations enables the interpretation of Young’s modulus and breaking strength of the 2D material, which are exhibited to strongly depend on the grain size. Introduction Pristine graphene with perfectly hexagonal carbon rings exhibits extraordinary mechanical properties. However, large-scale monolayers of graphene grown by chemical vapor deposition (CVD) usually contain topological defects and grain boundaries, which might strongly influence the apparent properties of the 2D material. There have been increasing experimental and simulation studies on the stiffness and strength of polycrystalline graphene. For example, atomic force microscopy (AFM) has been employed to measure the mechanical properties of monolayer-graphene in a nanoindentation setup, leading to the interpretation of 2D elastic modulus and breaking strength, respectively [1,2]. Both pristine and polycrystalline graphene (with varying grain sizes) have been measured by nanoindentation in AFM, and it was shown that the stiffness and failure load of poly-graphene are comparable with those of pristine graphene, despite the existence of grain boundaries [2]. In contrast, some other studies showed that grain boundaries severely reduce the elastic modulus and fracture strength of graphene [3,4]. For nanoindentation technique, the experimental measuremens and simulation predictions on the modulus and strength of polycrystalline graphene were found to be scattered widely [5-8], which were also different from those by uniaxial tension [9,10]. This work aims to explore the role of grain size in influencing the mechanical properities of polycrystalline graphene. Molecular dynamics (MD) simulations are employed to perform the nanoindentation test in mimicking the experimental measurements in AFM. Indentation of Polycrystalline Graphene: Atomistic Simulations In experiments, polycrystalline graphene is usually generated by chemical vapor deposition, in which the growth of graphene monolayers starts from multiple nucleation sites simultaneously. As such, the orientations of individual grains rely on the local nucleation sites and are randomly distributed [11]. To account for such growth, we perform computational procedure to generate a certain number N of randomly placed seeds within a square region, and use Voronoi tessellation method to divide the area into N polygons. Each polygon is filled by a pristine graphene fragment with a random orientation angle θi, which is represented by the angle between the armchair direction and the horizontal direction. Owing to the hexagonal honeycomb structure of graphene, the angle θi falls into the range between –π/6 and π/6. The atoms near grain boundary junctions that are closer than 1.4 Å to any 2nd Annual International Conference on Advanced Material Engineering (AME 2016) © 2016. The authors Published by Atlantis Press 215 neighboring grains are removed. This treatment is based on the observation that the system with grain boundary atoms closer than this value is unstable. In this way, the coordinates of all carbon atoms in polycrystalline graphene are generated within a simulation cell (selected as 20 nm by 20 nm). Fig. 1(a) shows the topological structure of a polycrystalline graphene sample with the mean grain size of 2 nm, which contains 127 grains in total. A circular region with radius R is selected from the sample for nanoindentation simulations, in which the atoms along the periphery are fixed as boundary condition (Fig. 1(b)). We confirm that the distribution of misorientation angle from this sample preparation resembles a random distribution, meaning that there is no preferred orientation for the grains (Fig. 1(c)). The distribution of grain size is Gaussian with the average value of 2 nm, as shown in Fig. 1(d). The polycrystalline graphene samples with varying grain sizes can be readily produced by changing the number N of nucleation sites within the given area. Fig. 1 (a) The topological structure of polycrystalline graphene with the average grain size of 2 nm. The simulation cell (20 nm by 20 nm) contains 127 grains that are randomly oriented. Colors in the figure are used to distinguish neighboring grains. (b) Nanoindentation simulation of a suspended polycrystalline graphene layer. The atoms along the periphery (in black) are fixed during the simulation, and the color bar represents the energy of the atoms. The distributions of (c) the misorientation angle and (d) the grain size of the sample in (a). Next, we use molecular dynamics simulations to study the nanoindentation behavior of such polycrystalline graphene. A circular membrane of polycrystalline graphene (Radius: R = 10 nm) is perpendicularly indented by a hemispherical diamond tip (Radius: r = 1 nm) at the membrane center, as shown in Fig. 1(b). Our nanoindentation simulations were performed at constant atom number (N), volume (V) and temperature (T) (i.e., NVT ensemble) by using the large-scale atomic/molecular massively parallel simulator (LAMMPS). The system temperature was maintained at 1 K by employing Nos’e-Hoover thermostat, and the initial velocities of atoms in each direction were extracted from Gaussian distribution for the given temperature. The adaptive intermolecular reactive empirical bond order (AIREBO) potential [12] was used for carbon-carbon interaction in the simulations, with a switch function parameter 2.0 Å for the breakage of C-C bonds [13]. For the atomic-scale interaction between the atoms from the indenter tip and graphene layer, respectively, Lennard-Jones potential was used to describe the van der Walls type interaction, namely,