Density Functional Theory Study Of Copper Agglomeration On Transition Metal Nitride Surface

The feature dimensions in integrated circuits are shrinking with the development of semiconductor industry. Design and development of new generation of interconnect materials that are more reliable with higher conductivity to replace the conventional aluminum-based interconnect is a subject of extensive research. Copper, in particular, has been widely accepted to be the preferred interconnect material due to its lower electrical resistivity and higher electromigration resistance than those of aluminum. However, the high diffusivity of copper in silicon and silicon dioxide was found to degrade or even destroy device performance. Therefore, an interlayer between copper and silicon substrate is introduced to prevent copper diffusion. Unfortunately, Cu films were found to agglomerate on diffusion barrier surfaces easily due to the poor adhesion between Cu and barriers, forming Cu islands on the surfaces that lead to decrease of conductivity. Fundamental understanding of Cu agglomeration processes on barrier surfaces is of essential importance to ensure successful applications of Cu interconnection.In this paper, we chose WN(001) surface as a barrier layer and performed a systematical study of Cu adsorption behavior on WN(001) surface. We firstly identified the energetically most favorable adsorption site of Cu on WN(001) surface by comparing the adsorption energies. We then calculated the dynamic behaviors of one Cu atom adsorb on WN surface by molecular dynamic simulation. The MD results and the potential energy surface indicate the high mobility and poor adhesion of Cu on the surface. Series calculations on the transition states of Cu agglomeration process and the charge distribution of Cu clusters on WN(001) surface were performed to understand the thermodynamics of Cu agglomeration. We show that copper clustering is a competitive process against atomic wetting and temperature is a critical factor of the agglomeration process. Two pathways were defined by simulating the growth processes of Cu cluster on WN(001) surface. The results indicate the temperature is also an important condition for growing. To prove our predication, a lot of ab initio molecular dynamics simulation study on Cu agglomeration on WN(001) surface at different temperatures were conducted. We found that the weak interaction between Cu and N is a main factor of Cu agglomeration and the temperature is a critical condition of agglomeration process. Lower temperature can decrease the Cu agglomeration. Our results provide a fundamental understanding of Cu agglomeration and a direction to solve the problems in semiconductor industry.