Kinetics of CVD graphene growth on polycrystalline copper and the influence of surface texture

Graphene growth kinetics have been studied with the aim of investigating two incompatible models which have been applied in the recent literature – a modified Johnson-Mehl-Avrami-Kolmogorov (JMAK) model and a modified Gompertz function. Graphene was grown by atmospheric pressure chemical vapour deposition (APCVD) at 1065°C on polycrystalline copper foil substrates and film growth as a function of time was studied by scanning electron microscopy (SEM). The graphene coverage with time was found to evolve sigmoidally, preventing a superficial differentiation between the two models. However, further analysis demonstrated that the modified JMAK model was incorrect due to a non-constant Avrami exponent and general incompatibility of the model assumptions with the physical nature of graphene growth by CVD. The Gompertz model was found to match the coverage data fairly well; however there are still some questions regarding the applicability of this model, the reasons for which are discussed. SEM micrographs demonstrate that certain copper grains supported selective nucleation and anisotropic growth of graphene domains at low densities, while nucleation was homogeneous and of fairly high density on other grains. These differences in nucleation and island morphology were amplified in samples grown at a lower temperature (1025°C), including formation of graphene domain chains on certain grains. Electron backscattering diffraction (EBSD) was employed to characterise the copper surface texture, which revealed that chains of islands form on step bunches present on very rough grains with orientations close to (111). Analysis of graphene coverage on different copper faces was carried out, which showed that the coverage remained approximately constant over the surface of a sample grown at high temperature, while growth at a lower temperature resulted in a decreasing coverage on grains with an increasing fraction of (111) terraces. Potential reasons for these orientation-dependent differences in graphene growth are discussed.