Electroless deposition offers the capability of metalizing non-conductive substrates and electrically isolated features. This is achieved by reducing the deposited ion using a chemical reducing agent rather than by externally provided current. However, despite its technological importance, the electroless process is still poorly characterized with little understanding of the factors that control the deposition rates. In particular, the significant effect of transport on the electroless process has received very little attention. The underlying model for electroless plating is the mixed potential theory, proposed by Wagner and Traud¹, which states that in order to assure charge balance, the electroless process must proceed at the rate and potential where the two separate redox reactions exhibit the same and opposite currents. While the basic stipulation of the mixed potential theory is always valid, its application to electroless systems has exhibited inconsistent results. This failure is most likely due to the interactions between the two redox chemistries, and effects due to the freshly deposited substrate in the complete system, which are not encountered in the separate half-cell processes. This leads to discrepancies between the polarization curves measured for the individual redox reactions and the prevailing conditions in the combined system, which are not accessible to direct electrochemical measurements. As detailed below, we address this deficiency using two different approaches. We apply our analysis to a system where the mixed potential theory fails to predict the observed rates.²The system consists of copper sulfate, ethylenediaminetetraacetic acid as a complexing agent, glyoxylic acid as a reducing agent, and sodium hydroxide which is used to adjust the pH to 12.8.

The first approach is based on polarizing the complete electroless system, allowing external current to flow simultaneously with the electroless process. From the difference between the measured amount of deposited copper and the total external current, the glyoxylic acid reaction rate can be determined. Translating the copper and glyoxylic acid reaction rates to currents, polarization curves can be generated for the two reactions (copper deposition and glyoxylic acid oxidation) as they occur in the full electroless system. As shown in Fig. 1, these rates are significantly different in the combined system as compared to the separately measured polarization curves. In particular, the copper deposition rate is significantly enhanced in the full electroless system as indicated in Fig. 1.

A second approach taken to analyze and model the electroless system is based on formulating rate expressions, in terms of reaction rate power law, for each of the two redox reactions taking place simultaneously in the electroless process. The rate expressions incorporate the reactants and products concentrations raised to arbitrary powers, and potential-dependent Arrhenius type rate constants. As both reactions occur in electroless plating on the same surface in the absence of external current, their rates must be equal. From this equality, separate expressions for current density and mixed potential are developed. The rate constants and powers are determined experimentally by seeking best-fit between the rate expression and measured electroless deposition rates across wide concentration ranges, resulting in equations 1 and 2.

The effect of transport was also incorporated in the model, as depletion near the surface changes the reactants concentration at the reaction site. To account for the transport dependence, electroless deposition experiments were performed on a rotating disk electrode (RDE). The model predicts that at lower rotation speed, both the surface concentrations of the reactants and the ensuing reaction rate will be lower. However, initially measured plating rates were significantly lower than predicted at the lower rotation rates. This was identified as being the consequence of partial surface blockage of the downward facing RDE by bubbles of hydrogen evolved in the reaction. Once the surface bubbles were eliminated, good agreement between the model predictions and actual measurements was noted, as shown in Fig. 2.

The final result is a predictive, semi-empirical model for both deposition rate and the mixed potential for electroless plating which accounts for reactant concentrations and transport effects. More generally, this study presents a framework for characterizing other electroless systems for better understanding, scaling, and control of the process.

Acknowledgements

Atotech GMBH is acknowledged for funding this study and for providing helpful input.

References

1. Wagner, C., & Traud, W. (1938). On the Interpretation of Corrosion Processes through the Superposition of Electrochemical Partial Processes and on the Potential of Mixed Electrodes. Zeitschriff fur Elektrochemie.
2. Yu, Lu, et al. "Autocatalysis during Electroless Copper Deposition using Glyoxylic Acid as Reducing Agent." Journal of the Electrochemmical Society (2013): D3004-D3008.