Molecular Dynamics simulations of thermal conductivity in composites consisting of aluminum oxide nanoparticles surrounded by polyethylene oxide

Molecular Dynamics (MD) simulations of heat flow in the composite systems consisting of aluminum oxide nanostructures surrounded by polyethylene oxide were performed using known forcefields with consistent treatment of covalent (polymer) and ionic (nanoparticles) components. A reverse non-equilibrium molecular dynamics (RNEMD) method [implemented in open source MD Simulator LAMMPS was utilized to impose a temperature gradient and obtain the values of thermal conductivity. Several simulation boxes containing layers (4 nm and 20 nm width) and spheres (3 nm and 6 nm radii) of aluminum oxide surrounded by polyethylene oxide have been built, equilibrated and subjected to RNEMD. The sizes of the boxes varied from 10/15 nm × 10/15 nm × 45/200 nm. The boxes contained 0.6*10⁶ to 3*10⁶ atoms. An enhancement of effective thermal conductivity from 0.3 W/m·K (for pure polymer) up to 1.1 W/m·K was achieved for the composites containing multiple 20 nm layers of aluminum oxide. The value of interfacial thermal resistance at the aluminum oxide/polymer interface obtained from the simulations was approximately 5*10^-9 m²K/W. Temperature profiles from RNEMD atomistic simulations were compared to known bulk models. Patterns of time averaged local heat flux in different components of the composite systems were calculated.