Molecular Dynamics Study of Thermal Energy Transport in Graphene-Based Materials

Self-heating is a critical issue in micro- and nano-electronics, and efficient heat removal is crucial for the long-term reliability and performance of micro- and nano-electronic devices. In many semiconductor devices, the power dissipation is non-uniform, which results in hot spots in the device and makes the heat removal very complicated. When two-dimensional materials are used for the electronic applications, heat dissipation problem can be aggregated because even a small amount of Joule heat can lead to a dramatic temperature rise in the ultrathin materials. In industrial practice, high-performance site-specific heat spreader and thermoelectric material based solid cooling component are two potential tools for the hot spot mitigation. Graphene, which has ultrahigh thermal conductivity, electrical conductivity and thermoelectric power factors, can potentially be used as the core material for these two applications. In my Ph.D. study, the thermal transport properties of graphene and its derivatives were investigated using classical molecular dynamics simulations. The research goal was to understand how the chemical functionalization, isotopic effects and structural modification to influence the phonon transport in graphene and across graphene-metal interfaces, which could provide important insights to the thermal transport physics in graphene, and are of practical significance for graphene-based devices in nano-electronics and thermal management applications.

Four projects have been finished in my Ph.D. study. In the first project, the phonon transport physics of pristine graphene, hydrogenated graphene and graphene oxide was investigated using large-scale molecular dynamics simulations. For the pristine graphene, highly ballistic thermal transport was observed. As for the hydrogenated graphene and graphene oxide, the thermal conductivity was significantly reduced when the hydrogen and oxygen coverage increased. For example, an oxygen coverage of 5% could reduce the graphene’s thermal conductivity by ~90%, and a coverage of 20% oxygen could lower it to ~8.8 W/mK. This value is even smaller than the calculated thermal conductivity of graphene in amorphous limit (~11.6 W/mK), which is usually regarded as the lower boundary of graphene’s thermal conductivity. The vibrational power spectral analyses showed that this large reduction in thermal conductivity was due to the significantly enhanced phonon scattering induced by the oxygen and hydrogen defects.

In the second project, the thermal conductivity of oxidized polycrystalline graphene was studied. Grain boundaries and the spontaneous oxidization around grain boundary regions are the inherent features of graphene. Our study found the thermal conductivity of oxidized polycrystalline graphene decreased as oxygen coverage increased, which was due to the phonon-defect scattering. However, the relative thermal conductivity reduction of oxidized polycrystalline graphene when the oxidization was localized at the grain boundary regions was much smaller compared to that of oxidized single-crystalline graphene with the same oxygen coverage. This was because the grain boundaries themselves are already strong phonon scatterers, such that oxygen atoms residing at the grain boundaries have far less impact on the thermal conductivity as compared to the oxygen atoms on uniformly oxidized single-crystalline graphene. An effective medium approximation model was also developed, which predicted that the influence of both grain boundaries and oxygen atoms on the thermal conductivity of oxidized polycrystalline graphene became smaller as the grain size increased, and the thermal conductivity of oxidized polycrystalline graphene approached that of pristine single-crystalline graphene at the large grain limit.

In the third project, the coherent and incoherent phonon transport behavior in ¹²C/¹³C graphene superlattices was studied. The existence of coherent phonons in superlattice structure was observed when the period length was small. By changing the period length of the superlattices and thus the interface density, a minimum thermal conductivity was observed, which implied the crossover from incoherent to coherent phonon transport. The thermal conductivity of the superlattices could be further decreased as we disrupted the coherence of phonons by manipulating and randomizing the superlattice structure. Our results showed that graphene – a two-dimensional material with intrinsically weak anharmonic phonon scattering – is an ideal platform for studying the nature of phonons.

In the last project, the interfacial thermal conductance of oxidized graphene /copper interface was studied. The cross-plane thermal conduction at graphene /copper interface is very poor due to the weak van der Waals interaction. Our study found that the oxidization could enhance the thermal conductance at graphene /copper interface due to the copper-oxygen covalent bonding. The time-domain thermoreflectance measurements revealed that the interfacial thermal conductance increased with the degree of graphene oxidization until a peak value was obtained at ~7.7% oxygen/carbon atom percentage. The molecular dynamics simulations verified that the strong interfacial covalent bonds were the key to the thermal conductance enhancement.

Overall, the projects conducted in my PhD study contributed to a better understanding of the thermal transport physics in pristine graphene, functionalized graphene, graphene superlattice and graphene/metal interface, which is of significant scientific importance. For application perspective, my studies offered valuable information for the design of graphene-based electronics and thermal management applications, such as graphene heat spreader and thermoelectric solid state cooling components.