Initial Stages of Direct Copper Electrodeposition: Simulation and Experiment

Copper electrodeposition is employed in the industry to provide electrical connection between the transistors in microchips. As the interconnect sizes between the transistors decrease, the problems related to methods of deposition seed layer arise. One way to eliminate these problems is to electrodeposit copper onto the candidate substrate directly. However poor nucleation of copper on the substrate remains a challenge. The current research focuses on the effect of additive chemistries in tackling these challenges. The influence of additives on the initial phases of copper electrodeposition on various substrates has been investigated experimentally. Also a theoretical model is developed to explain the impact of additives' interaction by the use of computer simulations.;Cu electrodeposition on thin film, highly resistive, Pt electrodes was studied experimentally. It is shown that nucleation influences thin film uniformity and that electroplating additives may also cause an increase or decrease in the non-uniformity due to their spatial variation in their acceleration or inhibition across the electrode.;The impact of organic additives and pulse-plating parameters on the initial stages of copper electrodeposition on ruthenium was characterized by the use of scanning electron microscopy. Results are compared for different additives for continuous plating and pulse plating conditions. Pulse plating has a significant impact on the nucleus density. Replacement of Cl- ions with Br- ions of the same concentration yields an increase in the nucleus density at the same current density.;In a more detailed study, the effect of polyethylene glycol (PEG) suppressor, on the nucleation of copper was investigated on three substrates: glassy carbon, as-received ruthenium and ruthenium that was pretreated by evolving hydrogen on its surface in diluted sulfuric acid. Scanning electron microscopy is used to characterize the effect of PEG concentration on nucleus density and the distribution of particle sizes. Results show that 300 ppm PEG increases the nucleus density by approximately a factor of six to eight on all three substrates over that obtained without PEG. Images suggest that progressive nucleation is observed on all surfaces.;A theoretical model was developed to explain and predict the effect of the suppressor molecules (e.g. polyethylene glycol and EO-PO-EO block copolymer) on copper nucleation. Numerical simulations were used to explore the impact of additives on chronoamperometric and galvanostatic behavior of the nucleation stage of a direct electrodeposition process, as well as their impact on particle size distributions.;In chronoamperometric studies, the rate at which the suppressor adsorbs onto the freshly nucleated metallic particles is a key parameter in determining the time at which a current maximum is expected and the value of the peak current density. Suppressors hinder the growth of particles and also potentially lower the barrier to nucleation, resulting in a particle size distribution with higher number of smaller particles in progressive nucleation.;In galvanostatic studies, a fast adsorption rate of the suppressor results in much higher nucleus densities. The simulations predict that nucleus density increase with EPE is greater than that for PEG, which is consistent with experiment. Further, the simulations are used to predict the impact of pulse plating parameters. The results showed that pulse plating waveform may be optimized to improve the particle density in the presence of suppressor molecules.