Investigation of Cu electroless deposition mechanism using OCP measurement with QCM for the application to Cu damascene process in semiconductor fabrication

Abstract

Semiconductor process has been dramatically developed by introducing Cu damascene process, which enables the replacement of Al to Cu as interconnection material. The development continues to improve the performance of devices, and it is achieved by scaling up of chip integration so far. However, recently, the reduction of Cu line width according to the integration makes it difficult to deposit both diffusion barrier and seed layer uniformly inside of narrow trenches and vias using conventional physical vapor deposition. Thus, alternative deposition methods are suggested to solve the problem, such as atomic layer deposition and electroless deposition. Cu electroless deposition attracts a lot of attention for the use of the next generation metallization method because the direct deposition is possible on both the conventional diffusion barriers and the next generation diffusion barriers such as Ru-alloy or Co-ally. In this study, the electrochemical real-time observation method for the Cu electroless deposition was designed: both the open-circuit potential and mass change of an electrode were measured simultaneously during the deposition. Using the method, the Cu electroless deposition could be investigated in detail, such as the mechanism of Cu electroless deposition and the adsorption behaviors of organic additives. Based on the study, the electroless bottom-up filling of sub-60 nm trenches was finally achieved, confirming the possibility that the Cu electroless deposition could be used for the next generation metallization method. In the investigation of Cu electroless deposition, the effect of each component in the electroless bath on Cu surface was understood preferentially. It was observed that the Cu was continuously oxidized in the alkaline bath. However, when ethylenediaminetetraacetic acid, a well-known complexing agent, was in the solution together, it was found to participate in the removal of Cu oxides formed on the surface as well as the complexation of Cu ions in the solution. Formaldehyde, a reducing agent, was adsorbed onto the Cu surface and inhibited further Cu oxidation. Both components maintained low oxygen content on the Cu surface in the alkaline solution. During the Cu electroless deposition process, the induction period was observed at the initial stage of the deposition and it was related with the time that the formaldehyde was adsorbed and became activated on the surface, indicating that the oxidation of formaldehyde was the rate-determining step. The effect of Cu oxides on the electroless Cu film was also investigated. It was revealed that the formaldehyde rarely adsorbed on the oxidized Cu surface unless the ethylenediaminetetraacetic acid removed the oxide by the complexation. The same phenomenon was observed in the real electroless deposition. The pre-formed Cu oxide caused the rough surface of electrolessly deposited film, resulting from the irregular adsorption of formaldehyde. The effect of additive was investigated by injecting additives during the deposition. Additives were polyethylene glycol (PEG), 2,2-dipyridyl, and 3-N,N-dimethylaminodithiocarbamoyl-1-propanesulfonic acid (DPS). The addition of PEG during the deposition caused the reduction of the deposition rate as the PEG was gradually adsorbed on the surface. It was revealed that the adsorption of PEG blocked active sites for the formaldehyde adsorption, resulting in the suppression effect. The adsorption kinetics of PEG was strongly dependent on the diffusion coefficient of PEG, which was directly related to the molecular weight. It followed the adsorption-controlled kinetics when the diffusion coefficient was high, whereas that of PEG with low diffusion coefficient showed the diffusion-controlled kinetics. The strength of the suppression effect was also affected by the molecular weight of PEG: increasing the molecular weight enhanced the suppression effect. The adsorption behaviors of 2,2-dipyridyl and DPS were also studied. 2,2-dipyridyl was found to adsorb on the surface fast and reduce the deposition rate immediately. DPS was also adsorbed, accelerating or inhibiting the deposition according to its concentration. It was revealed that the suppression effect was enhanced when both 2,2-dipyridyl and DPS were added together. The electroless filling was then performed on sub-60 nm trenches by using the suppression enhancement. The simple concentration optimization of two additives achieved the bottom-up filling.