(184c) Investigation of Roles of Chloride and Polyethylene Glycol In Copper Electrochemical Deposition

Along with the decrease in the critical dimension of semiconductor chips, there is a strong demand for higher electron mobility metal interconnects to permit faster signal transmission and reduce the resistor-capacitor (RC) delay of the integrated circuit (IC). In addition, the metal deposition process must be able to create void-free deposits in the chip interconnects, which currently are as narrow as 65 nm or smaller. Consequently, replacing the traditional aluminum interconnect with copper one has been shown to reduce the wiring resistance by as much as 45%. The void-free filling based on copper (Cu) electrochemical deposition (ECD) is often referred to as superfilling or superconformal deposition and is enabled by several additives in optimal concentration ranges and under proper processing conditions.

Chloride (Cl-) is the most common additive in copper plating chemistries. Affinity for adsorption of chloride onto copper surfaces has been studied using a variety of techniques. Polyethylene glycol (PEG) is used commonly in electroplating processes requiring the capability of through-hole plating or superfilling of submicron features. The adsorption of PEG on the copper surface requires the presence of chloride to act as the bridging ligand between PEG and the electrode surface. PEG as a result functions as an inhibitor or suppressor, which increases the overpotential for deposition relative to a PEG-free electrolyte. The ethylene oxide ligand, (-CH2-CH2-O-)/-EO-, is responsible for the unique properties of PEG.

In this study, we investigate individual roles of and interactions of PEG and chloride ion at the copper electrode surface in copper electrochemical deposition (ECD) from an acidic copper sulfate solution using electrochemical methods such as linear sweeping voltammetry (LSV), chronoamperometry, and AC electrochemical impedance spectroscopy (EIS). The LSV results reveal current inhibition in the presence of both PEG and chloride in the solution and role of chloride in copper ECD. The EIS results reveal a low frequency inductive loop associated with PEG adsorption, an intermediate capacitive loop associated with chloride bridge, and a high frequency capacitive loop related to double layer charging. A detailed reaction mechanism is proposed based on the experimental observations and an EIS model based on the reaction mechanism is developed. The EIS results are used to estimate parameters in a simple equivalent circuit model and the EIS model. The validity of the mechanism is established by comparing predictions of the EIS model with experimental EIS data. The mechanistic EIS model based on the reaction mechanism represents the experimental EIS data very well.