Effect of Current Density on Morphology of Lithium Electrodeposited in Ionic Liquid-Based Electrolytes

Fig. 1 shows the chemical formulae of the cations and anions of the ionic liquids and their abbreviations. [C₃mpip][Tf₂N], [C₄mpyr][Tf₂N], and [N_6,1,1,1][Tf₂N] were selected as the base electrolytes, because they are very stable at negative potentials such as the Li/Li⁺ potential^24,25 and their viscosities range between a certain extent. [C₃mpip][Tf₂N] and [C₄mpyr][Tf₂N] were purchased from Kanto Chemical Co. (Japan). Li[Tf₂N] and [N_6,1,1,1][Tf₂N] were purchased from Kishida Chemical Co. (Japan). Electrolyte preparation was carried out in dried air (dew point: below −40°C) after drying under vacuum at 75–85°C for 3 h. Li[Tf₂N] (0.32 mol kg⁻¹, 10 wt%) was added to [C₃mpip][Tf₂N], [C₄mpyr][Tf₂N], and [N_6,1,1,1][Tf₂N]. The residual water level in the resulting electrolytes, which was measured using a Karl-Fisher moisture meter (Kyoto Electronics Manufacturing, MKC-510N), was below 50 ppm. The viscosity of the electrolyte was measured using a cone-plate-type viscometer (Brookfield DV III+) in dry air at 25°C.

Zoom In Zoom Out Reset image size

A coin type two-electrode cell was fabricated in an Ar-filled glove box (dew point: below −80°C). The working electrode was nickel foil (Nilaco, Japan, 99%), and the counter electrode was lithium foil pasted on copper foil (Honjo Metal, Japan, Lithium battery grade). The distance between the two electrodes was fixed at 1 mm by a spacer ring. On the basis of Shiraishi et al.'s study,⁹ lithium electrodeposition was carried out at a total charge of 3 C cm⁻² (=3 A·s·cm⁻²). The current density was from 5 to 1000 μA cm⁻². After electrodeposition the cells were opened in the Ar-filled glove box. The working electrode was washed with dimethylcarbonate and transferred directly to the scanning electron microscope using a custom-made transfer vessel. The surface morphology of the working electrode was observed by SEM.

For the measurement of lithium-ion diffusion coefficient, a coin-type three-electrode cell was fabricated in an Ar-filled glove box. Both working and counter electrodes were lithium foil (φ14 mm). The reference electrodes consisted of a piece of lithium foil attached to the nickel lead. Glass fiber filters (Whatmann, GF/A, 260 μm thickness) were used as separators. The potential of the working electrode was set to −0.5 V, and current profiles were obtained. Although such a measurement should primarily be carried out using a large counter electrode, located far enough from the working electrode, a symmetrical cell with an electrode with a 260-μm-thick separator was used for this experiment in order to simulate a practical cell. To calculate lithium-ion diffusion coefficient, the current profiles were analyzed using the Cottrell equation,

where I(t)_lim is the limiting current, t is the time, n is the electrons per molecule oxidized or reduced to 1, F is Faraday's constant, A is the area of the electrode, c^* is the bulk concentration, and D₀ is the diffusion coefficient of lithium ion. In this experiment, n was 1 and A was 1.54 cm²; c^* was 0.46, 0.45, and 0.44 mol dm⁻³ for [C₃mpip][Tf₂N], [C₄mpyr][Tf₂N], and [N_6,1,1,1][Tf₂N], respectively. c^* was calculated from the density, which was measured using a pycnometer (AccuPyc II 1340, Micromeritics) in dry air at 25°C.

The voltage-coulomb amount profiles for the cells are shown in Fig. 2. In all cases, side reactions were observed when the voltage reached 1.5 V, which tended to consume more coulomb amount when the current density was small. The details of the side reactions are not well defined, but these reactions are thought to arise from the reduction of the native oxide on the substrate surface, or from the impurity in the electrolytes. For the cell with [C₃mpip][Tf₂N], the voltage rapidly dropped to below −4 V under a reduction current, with current densities of 500 and 1000 μA cm⁻² (Fig. 2(a1)). This rapid drop was due to the extremely low concentration of lithium ions in the vicinity of the working electrode.^26,27 Before the abovementioned voltage drop, two plateaus were observed at around −0.5 V and −2.0 V, which could be attributed to lithium deposition and a side reaction such as reduction of the ionic liquid, respectively. This rapid drop was also observed for the cell with [C₄mpyr][Tf₂N] and [N_6,1,1,1][Tf₂N], at current densities of 500 and 1000 μA cm⁻².

Zoom In Zoom Out Reset image size

When the current density was 200 μA cm⁻² or less, a plateau was observed at around 0 V for the three cells, considered to be due to lithium deposition. Fig. 3 shows the relationship between the plateau voltage and the corresponding current density of each cell. The difference between 0 V and the plateau voltage tended to increase when the current density was high and the electrolyte was more viscous. The plateau voltage is determined by the overpotential applied to the working electrode and the counter electrode. The decline of the plateau voltage from 0 V and the overpotential at the working electrode are proportional to each other, as the plating/stripping behavior obtained in cyclic voltammetry is symmetric. Note that the iR component derived from solution resistance is calculated to be very small related to the cell voltage, using measured ionic conductivity of the electrolytes (i.e. 0.84 mS cm⁻¹ for [C3mpip][Tf2N]-based electrolyte). More details on the relationship between the plateau voltage and the current density will be discussed later, taking into consideration the viscosity and diffusion coefficient data.

Zoom In Zoom Out Reset image size

Fig. 4 shows the SEM images of electrodeposited lithium on a nickel working electrode for the cell with three electrolytes of different current densities. Note that the observed coulomb amount for the cell with current densities of 500 and 1000 μA cm⁻² was smaller than 3 C cm⁻², which was the coulomb amount for the other cells. With an increase in current density, the size of the deposited particles decreased and the particles were more distributed. This can be explained as follows. Under the assumption that the lithium metal particles grown are semispherical in shape, the free energy change, ΔG, is expressed as

where ΔG_V is the free energy change per unit volume; V is the volume of the lithium metal semisphere; σ₁ and S₁ are the surface energy per unit area and the area of the interface between electrolyte and the lithium metal nuclei respectively; σ₂ and S₂ are the surface energy per unit area and the area at the interface between the nickel substrate and the lithium metal nuclei, respectively; σ₀ is the surface energy per unit area at the interface between the nickel substrate and the electrolyte.²⁸ ΔG_V has a negative value, −nFη, where n is the electrons per molecule oxidized or reduced, i.e., 1 in this case; F is Faraday's constant; and η is the overpotential. σ₀, σ₁ and σ₂ assume positive values. When the lithium metal nucleus grows to a semispherical shape,

where r is the radius of the lithium metal semisphere. ΔG is now expressed as

The critical radius for the nucleation, r^*, is determined by solving the equation dΔG/dr = 0

Thus, when a larger current density is passed, a higher overpotential η should also be applied, and the critical radius r^* should become smaller. The total charge amount for the reduction of lithium ions to lithium metal is constant, and in the process, the number of nuclei should increase. This explanation agrees well with the results shown in Fig. 4, illustrating the relationship between the current density and the number of nuclei.

Zoom In Zoom Out Reset image size

Next, the shape of the deposits is discussed. For the cell with [C₃mpip][Tf₂N], when the current density was 10 μA cm⁻² or less, circular deposits were observed. Circular deposits were also observed when the current densities were smaller than or equal to 10 μA cm⁻² and 50 μA cm⁻² for the cells with [N_6,1,1,1][Tf₂N] and [C₄mpyr][Tf₂N], respectively. For the cell with [C₃mpip][Tf₂N], when the current density was 50 μA cm⁻² or more (region between the dashed and solid lines in Fig. 4), needle-like deposits were observed, which are thought to have grown from semisphere-like nuclei. When the current densities were larger than or equal to 50 μA cm⁻² and 100 μA cm⁻² for the cells with [N_6,1,1,1][Tf₂N] and [C₄mpyr][Tf₂N] respectively, needle-like deposits were again observed, which were thought to have grown from semisphere-like nuclei. In general, noble metals such as copper and silver are electrochemically deposited in a dendritic form from an aqueous solution, under a rather larger current density. This is because the growth kinetics are diffusion-controlled in such a case. In the present study, lithium metal is also thought to be grown under diffusion-controlled kinetics, and hence, the dendritic form is observed when the current density is in the region denoted by the dashed line and solid line in Fig. 4. The dashed line indicates the boundary between diffusion-controlled and charge-transfer-controlled growth. The current density that corresponds to the dashed line is expected to be related to the diffusion coefficient of lithium ions in each electrolyte. Hence, the diffusion coefficients were measured (see the Experimental section). The diffusion coefficients were determined by using the Cottrell Equation (Eq. 1), and are summarized in Table I with the corresponding viscosity of each electrolyte. The diffusion coefficients are inversely proportional to the viscosity.

Fig. 5 illustrates the mechanism by which the current density determines the distribution of the lithium nuclei and the morphology of the lithium deposits. As for the distribution and the size, under a large current density, the nuclei are more distributed and the size of the nuclei is smaller. As for the morphology, when the current density is sufficiently small, the nuclei grow semispherically and the deposition becomes charge-transfer controlled. When the current density is sufficiently large, the nuclei grow in a dendritic form and the deposition is diffusion-controlled. When the current density is so large that the diffusion of lithium ions is slower than the lithium-ion reduction, the potential assumes a highly negative value and reactions other than lithium deposition become dominant.^26,27 When the nucleation and growth of deposits are considered individually, the distribution of nuclei and charge-transfer-controlled growth can be combined. Fig. 6a shows the SEM image of the lithium electrodeposited following the above concept under the following conditions: a current density of 1000 μA cm⁻² was applied for the first 0.15 C cm⁻², and a current density of 50 μA cm⁻² for the second 2.85 C cm⁻² (Fig. 6b). Rather small lithium deposits were observed, and the deposits were well distributed. This is one way to optimize the electrodeposition conditions, i.e., the charging conditions for morphology control.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The scheme shown in Fig. 5 is very common for the electrodeposition of noble metals from an aqueous solution, but it has not been reported for lithium metal. This is believed to be due to the reductive reactivity of lithium with ordinal electrolytes such as an organic solvent, in which case surface films are formed on the lithium metal. In the present study, the effect of the film formed by the reaction between the lithium metal and the electrolytes is thought to be small, as the selected ionic liquids are very stable at a negative potential such as the Li/Li⁺ potential. To suppress dendrite formation, the current density should be in the region of charge-transfer-controlled growth. This region can be controlled by changing the diffusion coefficient of lithium ions, which is affected by the solvent type, temperature, and lithium salt concentration.

The morphology of electrodeposited lithium in room-temperature ionic liquids was investigated by ex situ SEM observations, and the dependence of the nuclear distribution of the electrodeposited lithium on current density was discussed with relation to the lithium-ion diffusion coefficient. As schematically displayed in Fig. 5, the nuclei are smaller and more distributed when the current density is large. The nuclei grow in a semispherical shape in the charge-transfer-controlled mode when the current density is sufficiently small. The nuclei grow in the dendritic form in the diffusion-controlled mode when the current density is sufficiently large. When the current density is much larger so that diffusion of the lithium ions is slower than lithium-ion reduction, the potential becomes too negative and reactions other than lithium deposition occur. Considering that more nuclei are generated under a large current density and are grown at a small current density, small lithium particles are obtained and the deposits are well distributed. A method to control the current region with charge-transfer-controlled growth is presented.

This work was supported by JST-ALCA Project, "Next-generation Rechargeable Battery."