Multi-Scale Study of Branched Electrodeposits

Abstract

This study concerns the growth of metallic ramified branches formed by galvanostatic electrolysis of a stagnant metal salt aqueous solution inside a horizontal Hele-Shaw (thin gap) cell. Indeed, without supporting electrolyte, the electroneutrality constraint forces the deposit to grow rapidly in the form of long ramified branches made up of crystals whose size might be as lower than 100 nm [1-2] . The presentation focuses on the study of these branches growth, the understanding and the elucidation of the formation mechanisms of its nanostructure (crystals assembly) in order to exploit it as an alternative process for nanoparticles synthesis [2]. Besides, we propose here a new experimental protocol allowing to recover the whole electrodeposit assembly without damages, in order to process to its multiscale characterization and also the determination of the local microscopic morphology in each location of one branch. This presentation focuses specifically on the growth of the branches of copper, silver and iron at i) the macroscale for the branch pattern (10 µm to 1 mm, direct optical visualizations) and ii) also the scale of the nanocrystals (from 50 up to 500 nm, Scanning Electron Microscopy SEM). For each experiment, the branch pattern and structure are collected, and the measurements enabling correlating the impact of large scale transport phenomena on the resulting micro/nano-structure. The data are analysed with the support of theoretical models at both macro and micro scales. The effects of the current density J (16-266 mA/cm2), the electrolyte concentration C (0.1-0.5M), the type of metal (copper, silver, iron) and the cell thickness (25-100 µm) are investigated. At the macroscale, the experimental results are compared to the prediction of a specifically developed growth model based on a “2D dielectric breakdown” [4] scheme assuming a diffusion-limited process for the metallic cation (Mλ +) and taking into account a finite growth velocity vg for the electrodeposit. For low J/C, the experimental and theoretical patterns (see the figure) are fractal up to a cut-off length, corresponding to the diffusion length Ld = D/vg (D: diffusion coefficient). However, there are some differences between the theoretical and experimental apparent density of the deposit as well as in Sand time (time for the full depletion of the cations at the electrode marking the onset of the branch growth). These differences are related to non-diffusive mass flux generated by concentration gradients (natural convection) and the electric field (electroconvection) [5]. The resulting fluid flows have been observed by tracking the motion of polystyrene microparticles (15 µm) into the cell. Exacerbation of each one has been observed varying the cell depth (natural convection) and C (electroconvection). The obtained fluid velocity scales are used to discuss the effect of each of the non-diffusive flux on the branch pattern. At the microscale, the operating parameters affect both the size and the form of the crystals constituting the branches: two distinctive electrocrystallisation growth regimes of the deposit were distinguished, thanks to the SEM observations: one leading to the formation of non-dendritic crystals with well-developed facets and the other giving dendritic crystals (as shown in the figure). Indeed, for all precursors involved (Cu2+, Ag+ and Fe2+) at constant concentration for the deposition under high J, the branches are composed of dendrites, on which small non-dendritic crystals have nucleated and grown. The amount and the size of these non-dendritic crystals decrease when J increases. For Cu and Ag, and for the lower J values investigated, no dendritic structures are observed; the branches are only constituted by piles of non-dendritic crystals with rather wide sizes distribution. Dendrites are always observed in the case of Fe. Using the Fleury’s nucleation model [1] and considering the instability of Mullins Sekerka [6] adapted to the electrochemical case (providing a threshold for the onset of the growth instability, related to the onset of dendrite growth here), it is proposed that these microstructure differences are due to a competition between nucleation and out of equilibrium crystalline growth. This size distribution also differs according to the metal. This is related to their specific surface energies. [1] V. Fleury, Nature, vol. 390, no 6656, p. 145-148,Nov. 1997. [2] A. Iranzo and al, Electrochimica Acta, vol. 250, p. 348-358, Oct. 2017. [3 L. Niemeyer and al, Phys. Rev. Lett., vol. 52, no 12, p. 1033-1036, Mar 1984. [4] J. M. Huth, and al, Phys. Rev. E, vol. 51, no 4, p. 3444-3458, Apr. 1995. [5] W. W. Mullins et R. F. Sekerka, J. Appl. Phys., vol. 34, no 2, p. 323-329, Feb. 1963. Figure 1