Study of Dendritic Growth in Electrodeposition under Microgravity Conditions

We present in this paper a project aiming to study dendritic electrodeposition in zero-gravity conditions. Following previous experiments performed in a drop tower facility, this project will include parabolic flight experiments on the CNES/Novespace zero-g Airbus. The objectives of this project will comprise different aspects: i) a study in zero-gravity conditions of the nucleation phenomena involved in electrodeposition, ii) a better understanding of the interplay between the formation of a passivation layer on the surface of a lithium electrode and the onset of dendritic growth, iii) a quantitative test of a model proposed by J.-N. Chazalviel to explain the formation of irregular electrodeposits. The project will involve numerical calculations and experiments performed on lithium electrodes and other metals (copper, for example). It has been known for decades that metal electrodeposition from a binary electrolyte produces irregular deposits with various shapes. This phenomenon has attracted much interest in the last 50 years. A special interest has been devoted to the subject of dendritic deposition, which may be a serious problem in battery technology [1]. In particular this phenomenon has limited the use of all-solid-state lithium batteries with metallic lithium positive electrodes: although one of the most promising technologies to meet the requirements of electric power sources in novel applications (portable electronic devices or electric vehicles), this technology is not yet available because dendritic growth may appear during the charge of the battery (Fig.1). Dendrites may cause internal short-circuits. Also, the finely divided lithium is extremely reactive and thus may be the cause of serious fire hazards. Finally, most of the metal in the dendrites may become isolated, hence will not participate in further charge-discharge cycles. Fig.1 Dendritic growth in a lithium/polymer electrolyte visualisation cell: a current density of 0.05 mA.cm has been applied for 100 h. A dendrite has crossed the electrolyte, short-circuiting the cell (from Ref. [2]). Hence, avoiding these dendritic growth phenomena in Li batteries is still an important issue. However, for liquid electrolytes, and whatever the electrochemical system under concern, a precise study of dendritic growth mechanisms may be complicated by buoyancy driven convective motion [3]: even in thin, quasi-two-dimensional horizontal cells, electrodeposition is accompanied by a gravity-induced fluid flow at the electrodes (Fig.2). This effect is due to the electrolyte density decrease at the cathode (due to the concentration decrease) or increase at the anode (concentration increase): it has been extensively studied in the recent literature, both theoretically [4-6] and experimentally [7-11]. Convective motion mixes the electrolyte and tends to homogeneise the concentration. The effect depends on cell thickness, salt concentration, and current density [4]. Gravity-induced convection was shown to increase the instability of the system [11]. Obviously, zero-gravity experiments [12] should give a straightforward mean to avoid this effect: apart from gravity, no other parameter is altered. However, these experiments are relatively difficult to work out and of short duration (around 5 25 s in easily available facilities). Hence, several other techniques have been proposed: a) very thin cells (cell thickness less than 30 μm) and/or electrolytes with very low concentration, below 0.01 mol L-1 [9]: however this concentration range is not compatible with concentrations used in actual Li batteries. b) gel [7], or electrolytes with increased viscosity [13] Fig.2 Schematic view of an electrochemical cell in the vicinity of the electrodes: a convective motion is induced in these regions, due to the electrolyte density decrease at the cathode (concentration decrease) or increase at the anode (concentration increase). allowed for measurements without the above limitation. However, the chemistry involved in these electrolytes may be markedly different from that involved in liquid solutions. c) thin cells in a vertical position, with cathode set horizontally on top of the cell, were used by several authors [7, 9, 11, 14-17]. For example, Marshall et al. [11] showed that this configuration enables to obtain a stable regime. Using this method, Morisue et al. could evidence the effect of gravitation on nucleation phenomena in copper electrodeposition onto a TiN substrate [18]. Gonzalez et al. could measure impedance diagrams of the electrolytic cell at the onset of dendritic growth [17]. These experiments, with a cathode set horizontally on top of the cell, are easy to handle, and they have no time limitation. However, obviously gravity is still present, and convective motion may appear, in particular when the current density distribution (hence the concentration distribution) is not uniform along the electrodes. In conclusion, it appears that avoiding buoyancy driven convection in electrodeposition requires zero-gravity experiments when one wishes to avoid the above mentioned parasitic effects. Several studies carried out by the Department of Energy Science & Technology (DEST) of the Kyoto University have shown that even short zero-gravity experiments may provide interesting information on dendritic growth in electrodeposition. In particular the DEST performed copper electrodeposition experiments from CuSO4 aqueous solution [12]. The experiments were carried out in a drop shaft allowing free falls for more than 8 seconds under zero-gravity environment. Experiments were performed at constant current density. Ionic concentration in the electrolyte was measured from interferometry: the diffusion layer thickness was shown to increase as the square root of time, in good agreement with theory. Due to natural convection the diffusion layer thickness was shown to increase much faster at 1g than at 0g (Fig.3). Also, at 0g, the deposit consisted in fewer, larger grains than at 1g. Fig.3 Time variation (as a function of square root of time) of the diffusion layer thickness measured in copper electrodeposition from copper sulfate in 1g and 0g conditions (from Ref. [12]). Another experiment carried out in the drop shaft facility consisted in measuring concentration profiles around Ag dendrites electrodeposited in 3M AgNO3 aqueous solution [19]. Both the DEST and Laboratoire de Physique de la Matiere Condensee of Ecole Polytechnique are now working on a joint project aiming to study electrodeposition in the zero-gravity environment provided by parabolic flights. This project, is sustained by the Centre National d'Etudes Spatiales (CNES): it plans electrodeposition experiments on lithium electrodes and other metals (copper, for example) in the Airbus A300 from CNES/Novespace. During these flights, almost zero-gravity conditions are available over 20 to 25 seconds. The results of these experiments will be compared with those of longer experiments performed in vertical cells. Numerical calculations will also be held, in particular to obtain ionic concentration distributions in the electrolyte. The objectives of this project are the following: a study of nucleation phenomena on the electrodes: in particular we wish to extend in zero-gravity conditions the study of nucleation of copper on a TiN electrode performed by Morisue et al. [18]; a better understanding of the interplay between the onset of dendritic growth and the formation of a passivation layer on the surface of a lithium electrode. In particular, we wish to study, in zero-gravity conditions, the incubation period which was observed in the formation of concentration gradients at the initial stage of electrodeposition experiments [16, 20]. a quantitative test of a model proposed by J.-N. Chazalviel [21] to explain the formation of dendritic deposits in electrodeposition. This model predicted the formation of a non classical space charge in the vicinity of the electrode: it was never verified experimentally.