Improving Copper Electrodeposition in the Microelectronics Industry

Sporadic voiding within the interfacial Cu3Sn intermetallic compound (IMC) layer-sometimes referred to as "Kirkendall voiding"-has been found to lead to degradation of solder joint reliability in board level, mechanical shock testing. It has been suggested that the voiding phenomenon is a result of the incorporation of organic impurities in the copper (Cu) deposit during electroplating. In the present study, Cu samples were electroplated galvanostatically from a generic solution, containing Cl- ions, as well as a suppressor [polyethylene glycol (PEG)], and a brightener (bis(3-sulfopropyl) disulfide, SPS) as additives. Overpotential transients were measured during electroplating with a range of current densities (0.5-40 mA cm(-2)) in baths with various compositions. Effects of the bath chemistry on the Cu surface morphology, as well as on the propensity for voiding after soldering, were also investigated. Elemental analysis of selected samples was performed by SIMS. Plating at 10-20 mA . cm(-2) with an optimized bath composition led to Cu with a fine-grain structure and smooth appearance. Solder joints formed from these deposits remained void free after soldering and thermal aging. Lower current densities, ran in the same plating bath, led to a significant propensity for voiding, apparently because of incorporation of, principally, SPS and its breakdown products into the growing layer. Continuous plating at 10 mA cm(-2) for up to 18 hours without replenishment revealed a strong dependence on bath aging, with Cu changing from "void-proof" to clearly "void-prone." These trends were attributed to the different rates of consumption for PEG and SPS and changes in the contaminants being incorporated in the deposits. In general, differences in the voiding behavior of the plated Cu could be predicted by monitoring a set of characteristic overpotential transient signatures.