Indexed on: 16 Mar '16Published on: 16 Mar '16Published in: JOM (Warrendale, Pa. : 1989)
The Green–Kubo method was used to investigate the thermal conductivity as a function of temperature for epoxy/single wall carbon nanotube (SWNT) nanocomposites. An epoxy network of DGEBA-DDS was built using the ‘dendrimer’ growth approach, and conductivity was computed by taking into account long-range Coulombic forces via a k-space approach. Thermal conductivity was calculated in the direction perpendicular to, and along the SWNT axis for functionalized and pristine SWNT/epoxy nanocomposites. Inefficient phonon transport at the ends of nanotubes is an important factor in the thermal conductivity of the nanocomposites, and for this reason discontinuous nanotubes were modeled in addition to long nanotubes. The thermal conductivity of the long, pristine SWNT/epoxy system is equivalent to that of an isolated SWNT along its axis, but there was a 27% reduction perpendicular to the nanotube axis. The functionalized, long SWNT/epoxy system had a very large increase in thermal conductivity along the nanotube axis (~700%), as well as the directions perpendicular to the nanotube (64%). The discontinuous nanotubes displayed an increased thermal conductivity along the SWNT axis compared to neat epoxy (103–115% for the pristine SWNT/epoxy, and 91–103% for functionalized SWNT/epoxy system). The functionalized system also showed a 42% improvement perpendicular to the nanotube, while the pristine SWNT/epoxy system had no improvement over epoxy. The thermal conductivity tensor is averaged over all possible orientations to see the effects of randomly orientated nanotubes, and allow for experimental comparison. Excellent agreement is seen for the discontinuous, pristine SWNT/epoxy nanocomposite. These simulations demonstrate there exists a threshold of the SWNT length where the best improvement for a composite system with randomly oriented nanotubes would transition from pristine SWNTs to functionalized SWNTs.