The contact conductance of carbon nanotube (CNT) junctions is the key factor that controls the collective heat transfer through CNT networks or CNT-based materials. An improved understanding of the dependence of the intertube conductance on the contact structure and local environment is needed for predictive computational modeling or theoretical description of the effective thermal conductivity of CNT materials. To investigate the effect of local structure on the thermal conductance across CNT-CNT contact regions, nonequilibrium molecular dynamics (MD) simulations are performed for different intertube contact configurations (parallel fully or partially overlapping CNTs and CNTs crossing each other at different angles) and local structural environments characteristic of CNT network materials. The results of MD simulations predict a stronger CNT length dependence present over a broader range of lengths than has been previously reported and suggest that the effect of neighboring junctions on the conductance of CNT-CNT junctions is weak and only present when the CNTs that make up the junctions are within the range of direct van der Waals interaction with each other. A detailed analysis of the results obtained for a diverse range of intertube contact configurations reveals a nonlinear dependence of the conductance on the contact area (or number of interatomic intertube interactions) and suggests larger contributions to the conductance from areas of the contact where the density of interatomic intertube interactions is smaller. An empirical relation accounting for these observations and expressing the conductance of an arbitrary contact configuration through the total number of interatomic intertube interactions and the average number of interatomic intertube interactions per atom in the contact region is proposed. The empirical relation is found to provide a good quantitative description of the contact conductance for various CNT configurations investigated in the MD simulations and is suitable for incorporation into mesoscopic models capable of predicting the effective thermal transport properties of CNT materials.

• Received 23 December 2015
• Revised 1 June 2016

DOI:https://doi.org/10.1103/PhysRevB.94.014308