Connected and automated vehicles (CAVs) have the potential to revolutionize the transportation system by improving traffic conditions, increasing road capacities and reducing congestion and crash risks. This study aims to investigate the implications of this technology on greenhouse gas (GHG) emissions levels through traffic micro-simulations and emissions modelling. Three different driving behaviours (cautious, normal and aggressive) of CAVs are simulated along with a base driving behaviour of driver-operated vehicle (DOV) with varying traffic demands and under different network configurations to evaluate CAVs' implications. Results of this study show that aggressive CAVs can reduce large percentages of emissions compared to DOVs but can also increase emissions at situations of increased traffic demand and high congestion due to their higher acceleration rates which can cause irregular driving cycles. Analysis of the findings suggests there are optimal traffic demand levels of aggressive CAVs that can maximize the reduction of emissions relative to DOVs.