Electrodeposition, particularly when carried under transport limitations, may result in rough surface texture. The roughness elements often evolve into dendrites which may cause internal shorts in batteries, leading to catastrophic failure. Since just about all commercial electroplating processes use additives to control the texture and distribution of electrodeposits, it has been suggested to use similar additives to eliminate or minimize dendritic growth.

Strickler and Landau presented an analysis¹ showing that if inhibiting additives were to preferentially adsorb on the tips of the roughness elements, dendritic growth could indeed be suppressed. However, these authors also argued that any additive that adsorbs and accumulates on the surface will eventually cover the entire electrode, eliminating the preferential inhibition at the tips and thus negating the beneficial effects. In order to maintain the preferential tip coverage, additive removal mechanism must be provided, e.g., by additive incorporation into the deposit. A correct balance of additive transport, adsorption, and incorporation could maintain the tips preferentially covered while the remainder of the electrode can stay essentially free of inhibiting additives, leading to level deposition¹.

We report here the results of an experimental study that was conducted in order to test this hypothesis. Although copper is not used in advanced batteries, it was selected as a model system because the interaction of different additives in the copper system is well-characterized^2,3. Copper was plated from acidified copper sulfate (0.1 M) solution onto a rotating disk electrode, in the presence of different additives. In order to accelerate roughness evolution and increase the number of roughness elements to yield reliable statistics, deposition was conducted close to the limiting current (i/i_L=0.95). Following deposition (209 coulomb/cm²), the height of the roughness elements (0-200 mm) was measured by focusing on the roughness elements tips an optical microscope objective with calibrated stage elevation.

Two additives were compared: polyethylene glycol (PEG) and polyethyleneimine (PEI). PEG adsorbs weakly on copper (through its polyether oxygen) and is known to remain on the electrode. By contrast, the stronger adsorbing PEI (due to its nitrogen moiety) is incorporated within the deposit.

Results, summarized in Fig. 1, indicate, as expected, only small reduction in number and height of the roughness elements in the presence of PEG as compared to the pure copper system. Two PEG systems were compared: PEG 4000 (100 ppm) by itself, and PEG 4000 (100 ppm) in the presence of chloride (70 ppm), which is known to enhance the PEG adsorption². Both PEG systems reduced the number of large (150 – 200 mm) roughness elements by only about 14% as compared to the pure copper system. By contrast, the PEI (10 ppm), which is known to incorporate within the deposit, reduced the number of large roughness elements by a significant amount (42%).

These results give credence to the hypothesis that incorporating additives should be effective in inhibiting roughness evolution and dendritic growth, while additives that remain on the electrode are expected to have little effect. However, in order to irrevocably validate the hypothesis, more studies, encompassing additional additives, with precise control of the additives transport and incorporation balance, should be conducted.

Acknowledgements

This research was conducted as part of the NSF/DOD funded Research Experience for Undergraduates (REU) program on Electrochemical Engineering at Case Western Reserve University. NSF/DOD stipend to K.G. is gratefully acknowledged.

References

1. Alaina Strickler and Uziel Landau, “Arresting Dendritic growth during electrodeposition using special additives”, Abstract # 993, 223^rd, Electrochemical Society Meeting, Toronto, Canada, May 15, 2013.
2. Rohan Akolkar and Uziel Landau, Electrochem. Soc., 151, C702 (2004).
3. Julie Mendez, Rohan Akolkar and Uziel Landau, Electrochem. Soc. 156, (11), D474-D479 (2009).