An acid-copper plating bath containing chloride ions, polyethylene glycol ͑PEG͒, bis͑3-sulfopropyl͒ disulfide ͑SPS͒, and Janus Green B ͑JGB͒ has been characterized by electrochemical methods and by fill studies on electrode surfaces patterned with trenches ranging in size from 200 to 600 nm, with aspect ratios between 2 and 4. The electrochemical methods employed include linear sweep voltammetry ͑LSV͒, chronoamperometry ͑CA͒, and cyclic voltammetric stripping ͑CVS͒ on a rotating disk electrode. Comparison of the methods shows that in all the cases an analysis of LSV and CA gives consistent estimates of an effective surface coverage of additives. CVS results are consistent with the other results only in baths that are free of JGB. The ability to produce void-free deposits of a bath containing all of the additives can be understood in part with LSV and CA results. However a very good superfilling can also be obtained from a plating bath containing chloride ions, PEG, SPS, but no JGB. This observation cannot be easily interpreted with measurements within the context of established leveling theories.Commercial acid-copper plating baths may contain many organic additives. 1-4 Detailed knowledge of the function of each additive, as well as the interactions between these additives, is lacking despite the advanced state of the technology. The complexity of the bath chemistry, coupled with the incomplete understanding of the mechanisms, makes difficult theoretical predictions of the efficacy of an electrolyte with an arbitrary composition. Industry requires the development of experimental protocols for investigating the influence of composition of perhaps every component of an additive cocktail. The screening method should be fast, reproducible, easy to operate, and with easily interpretable results.We discuss possible screening methods as applied to an acidcopper electrolyte with chloride ions, polyethylene glycol ͑PEG͒, bis͑3-sulfopropyl͒ disulfide ͑SPS͒, and Janus Green B ͑JGB͒. The leveling capability of submicrometer, patterned trenches by this electrolyte has been reported. 1,2 In this paper, we characterize this system by cyclic voltammetric stripping ͑CVS͒, linear sweep voltammetry ͑LSV͒, and chronoamperometry ͑CA͒ studies on a ͑nonpatterned͒ rotating disk electrode ͑RDE͒. Additionally, we investigate the influence of current density, additive composition, and the fluid-flow conditions on the superfilling effect. We show that the LSV and CA methods can sometimes provide a consistent explanation for trench filling experiments; in a companion article in this issue, 5 we show the measurements can be interpreted quantitatively to provide thermodynamic, mass transfer, and kinetic parameters for use in shape-change simulations.
ExperimentalFill studies.-The pretreatment of the patterned wafer as well as the deposition procedure is described in Ref. 2. In the present study, copper deposition was carried out under a well-mixed condition, instead of in a stagnant solution. The wafer fragment was placed inside an RDE adapte...