Files
Abstract
Graphene is a layer of covalently bonded carbon atoms with a honeycomb lattice structure. It has exceptional properties, due to which it became prominent in a wide range of applications. The enormous success of graphene raised curtains to a new class of two-dimensional materials. In this research we employed molecular dynamics simulations and machine learning methods to study thermo-mechanical properties of graphene-like two dimensional materials. Reverse non-equilibrium molecular dynamics method has been employed to study the thermal transport properties of hexagonal-boron nitride (h-BN) nanoribbons, and C3N nanotubes (C3NNT). Our results showed that effective thermal conductivity of h-BN nanoribbons is in the range of 75 - 160 W/m-K whereas Kapitza conductance of grain boundaries is in the range of 6 - 20 GW/m2-K. By increasing the misorientation angle the defect density at the grain boundaries increases due to which, the Kapitza conductance of the grain boundaries and the effective thermal conductivity of the h-BN nanoribbon decreases. Our MD simulations showed that a lower ballistic to diffusive transition length (72 - 80 nm) for C3NNT compared to CNT (103 - 107 nm). Due to the stiffer acoustic modes of CNT. Thermal conductivity of CNT is significantly higher than that of C3NNT across the entire ballistic-diffusive range. Molecular dynamics simulations are used to study the mechanical properties of graphene-like two-dimensional material with focus on MoS2 and graphene. The toughness and strength of the MoS2 sheets are not significantly affected by increase in the number of layers from one to three. Griffith theory is not valid for nanoscale cracks of MoS2. In comparison to Inglis theory, quantized fracture mechanics give better prediction for the fracture strength of MoS2 sheets when the crack tips are very sharp. Environmental molecules like H2, O2, CO2, and H2O can react with material, and accelerate its failure. Hence, it is important to understand the environment-assisted cracking in two–dimensional materials. The stress corrosion cracking in mono crystalline graphene in the presence of O2 molecules are studied. The strained graphene sheet is exposed to oxygen molecules. Our MD and density functional-based tight bonding simulations show that cracks in graphene can grow due to chemical reactions with environmental molecules. The results show that sub-critical crack growth can occur in graphene sheets when exposed to O2 molecules. The mechanical and fracture properties of bicrystalline and polycrystalline graphene sheets with hydrogenated grain boundaries are studied. Molecular dynamics simulations are used to extract the traction-separation laws (TSL) of the hydrogenated grain boundaries of bicrystalline graphene. The adsorption site is an important factor for determining the level of the impact of hydrogenation on the fracture properties of the grain boundary. On the other hand, crack propagation path in the polycrystalline graphene sheet is affected by the hydrogenation of the grain boundaries. The crack prefers to grow along an intragranular path at lower hydrogenation percentage, whereas at higher hydrogenation percentage it changes to intergranular. Our results have showed that hydrogen embrittlement at grain boundaries is degrading the bicrystalline and polycrystalline graphene sheet due to which the strength and fracture toughness is decreasing. Finally, A machine learning model has been developed to predict crack propagation path in polycrystalline graphene. The model is a combination of convolutional neural network, bidirectional recurrent neural network, and fully connected layers. The data set used to train the machine learning model is obtained using MD simulations. Fully trained ML model can predict the crack path in polycrystalline graphene sheet. The predicted crack path from the ML model is in close agreement with the one obtained from MD simulations.