Additives and nucleation and growth in copper electrodeposition

The ability to predict the limitations of electrodeposition when extended to sub-100-nm features depends strongly on an understanding of the modes of action of additives. Nonetheless, very little fundamental understanding of additives during electrodeposition exists. In addition, at such a feature size, due to sub-optimal sidewall coverage of seed layers, the direct electrodeposition of copper onto barrier materials may be required. Thus, the present research focuses on additives and nucleation and growth in copper electrodeposition.; An acid-copper plating bath containing chloride ions, polyethylene glycol (PEG), bis (3-sulfopropyl) disulfide (SPS), and Janus Green B (JGB) has been characterized by electrochemical methods which include linear sweep voltammetry (LSV), chronoamperometry (CA), and cyclic voltammetric stripping (CVS) on a rotating disk electrode and by fill studies on electrode surfaces patterned with trenches. 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 which give consistent estimates of an effective surface coverage of additives in all the cases. CVS results agree well with the other results only in baths that are free of JGB. A model describing the effect of an accelerator (SPS) and two inhibitors (PEG and JGB) on the leveling efficacy on submicron trenches is simulated and compared to experimental fill studies. The simulations are consistent with fill studies over a large range of SPS and JGB concentrations. However, that a very good superfilling is also obtainable from a plating bath containing chloride ions, PEG, SPS, but no JGB is neither easily interpreted with the established leveling theories nor predicted by simulations. Compounds related to SPS in acid-copper plating baths are also tested for their electrochemical activity using CVS on a rotating disk electrode.; We have also explored the possibility of direct electrodeposition on Ta and TaN barriers, evaluating two electrolytes: copper-EDTA and copper-citrate baths. Effect of solution composition, pH, potential difference, current density, deposition time, as well as adhesion on nucleation and growth are discussed.; A nucleation and growth model assuming a concentration-dependent nucleation rate is simulated and compared to experimental results. Simulations predict an experimentally observed, but not previously explained, saturation of nuclei near the time of the current maximum. The impact of relaxing the usual assumption that nucleation rate is independent of reactant concentration on particle statistics is discussed.