Aqueous Zinc Sulfate Flow Through a Copper Mesh Anode Improves Zinc Metal Electrodeposition Morphology and Impedance

In the LSV tests, a platinum (Pt) wire reference electrode (99.95%, Thermo Fisher Scientific) is placed 2 mm away from the Cu wire mesh. The voltage is swept with a 5 mV s⁻¹ scan rate. Figure 2 shows the measured current density versus applied voltage at (a) low, (b) intermediate, and (c) high ZnSO₄ concentrations. In the low polarization regime, a linear relationship between applied voltage and current indicates that the electrolyte in the vicinity of the electrode replenishes the consumed zinc ions. When the current density increases to a critical value, the curve slope changes. Further polarization induces hydrodynamic instability indicated by random and irregular current responses. ³⁷ The slopes of current density versus overpotential change at −1.45 V, −1.45 V and −1.65 V versus Pt. in 0.01 M, 0.5 M and 1 M ZnSO₄ electrolytes, respectively, indicating that the impedance changes with increased polarization. The changed impedance at large polarization is possibly due to HER and electroconvection. ³⁷ Here, we choose the current density at the transition point of the slopes as the reference current density in stagnant electrolytes (${i}_{{RS}}$). The corresponding (${i}_{{RS}}$) are 5, 77, and 306 mA cm⁻² in 0.01 M, 0.5 M and 1 M electrolytes, respectively. In 0.01 M electrolyte, the reference current densities with flow (${i}_{{RF}}$) increases to 7 and 10 mA cm⁻² at Pe = 2, 4, respectively. Changed slopes of current density versus overpotential are not clearly observed for 0.5 and 1 M electrolyte. ${i}_{{RF}}\gt {i}_{{RS}}\,$ indicates that flow replenishes zinc ions and delays the onset of electroconvective instability. Flow increases current at the same voltage, so impedance is reduced.

Zoom In Zoom Out Reset image size

The flow velocity distribution is not uniform in the cell, with the highest velocity at the center of the chamber and zero velocity at the chamber walls. This results in different Zn electrodeposition morphology as shown in Fig. s2. To achieve uniform flow across the electrode, a flow distributor (e.g., quartz particles ³⁴ ) or an electrode with a low permeability (e.g., carbon paper ³⁸ or carbon strip ³⁹ ) can be used. Low permeability, however, requires higher pressure and more pumping power, reducing the efficiency of the battery. The trade-off between uniform flow velocity, pressure, and pumping power will be investigated in our future work.

In this work, we examine the center of the wire mesh in stagnant and flowing electrolytes to compare electrodeposition morphology. The center is approximately 0.008 mm² or 1/36 of the tube cross section, so the flow velocity is relatively uniform. The wire mesh is very porous, so the pressure drop is negligible.

Zn electrodeposition morphology in galvanostatic tests for 0.5 M and 1 M electrolyte are presented in Figs. 3a–3f and 3g–3l, respectively. Two galvanostatic tests are performed at the same current density of 47.4 mA cm⁻² for 0.5 ($0.62{i}_{{RS}}$) and 1 M ($0.15{i}_{{RS}}$) electrolyte, respectively. Figures 3a and 3g show that, in stagnant electrolytes, the deposited zinc morphology is bulbous and has randomly oriented plate-like structures. Under constant current, the ion concentration gradient near the zinc anode increases with time, enhancing zinc electrodeposition in protruding areas. The accumulated zinc deposits gradually form bulbous clusters that could fill the interelectrode space in a zinc-ion battery, causing short circuits. In contrast, no obvious zinc protrusions are observed in the flow-through electrolyte (Figs. 3d and 3j), indicating that the flowing electrolyte may homogenize ion concentration.

Zoom In Zoom Out Reset image size

To investigate the impact of flow on morphology below ${i}_{{RS}}$ for 0.5 M electrolyte, we choose 20.3 mA cm⁻² (${0.26i}_{{RS}}$) and 45.0 mA cm⁻² ($0.58{i}_{{RS}}$) as applied current densities. The flow rates are based on the critical Peclet number ${{Pe}}_{{cr}}\,$ at which advective transport rate equals diffusive transport rate at the electrode surface, or ${{Pe}}_{{cr}}=1.$ For flow rates above ${{Pe}}_{{cr}},$ convection controls mass transport. Figures 3b, 3c, 3e, 3f show the positive influence of electrolyte flow on zinc electrodeposition for 0.5 M electrolyte. The bulbous structures in Figs. 3b, 3c under no flow conditions are replaced by smooth deposits with smaller-size clusters in Figs. 3e, 3f. For the same flow rate, bulbous structures can be observed again in high current densities ($0.58{i}_{{RS}}$) in Fig. 3f, suggesting that higher flow rates may be needed for high power. We observe similar phenomenon in 1 M electrolyte where flow improves Zn electrodeposition morphology at 32 mA cm⁻² ($0.10{i}_{{RS}}$) in Figs. 3h, 3k and 71.1 mA cm⁻² ($0.23{i}_{{RS}}$) in Figs. 3i, 3l. We note that even at relatively high flow rates (e.g., ${Pe}=10$), the flow velocity is on the order of millimeters per second, easily within range of many low flow rate and compact (e.g., peristaltic) pumps. The power required is also expected to be small because low flow rates will produce small back pressure if the electrodes are sufficiently porous. Optical microscope photos of Zn electrodeposition morphology are shown in Fig. s3.

Table s1 shows that plating current densities and capacities of Zn electrodeposition in ZnSO₄ electrolyte with concentrations from 1 M to 3 M mostly range from 1 to 20 mA cm⁻² and 1 to 20 mAh cm⁻², respectively. Table s2 shows that, in redox flow batteries with different chemistries, the applied current densities of Zn electrodeposition mostly range from 20 to 80 mA cm⁻² with capacities from 10 to 320 mAh cm⁻². Large electrodeposition capacity (>100 mAh cm⁻²) may significantly shorten the flow batteries lifespan from hundreds to less than ten cycles. ⁴⁰ The electrochemical energy is stored in the electrolyte in flow batteries and greatly determined by the volume of the stored electrolyte. The flow-through electrolyte studied in this work is compatible with most conventional batteries with metal where the electrodeposition capacity is mainly determined by the electrodes. Non-dendritic Zn morphology with 71.1 mA cm⁻² and 15.8 mAh cm⁻² capacity in this work indicates that flow-through electrolyte can improve the surface stability at large polarization and high-areal capacity relative to the state-of-the-art.

Figures 4a–4c show mesh cross sections for 0.5 M stagnant electrolyte, and Figs. 4d–4f exhibit cross-section morphology at Pe = 10. Dendritic structures are observed on the surface closer to the zinc source in the stagnant electrolyte (Fig. 4a), but not in the flowing electrolyte (Fig. 4d). The dendrites formed in stagnant electrolyte are needle-shaped structures with approximately 128.5 μm maximum height in Fig. 4b. In contrast, the electrode shows flat metal deposition at Pe = 10 in Fig. 4e with small, uniform structures. Most zinc deposits at the front side of the wire mesh closer to the zinc source. The deposition thicknesses on the front and back side in Figs. 4c and 4f are presented in Table I. Due to transport limitations, zinc ions preferentially deposit on the front side of the wire mesh with the shortest path to the zinc source, while fewer zinc ions take the long route to the back side. Flow helps distribute Zn plating more uniformly around the wire diameter, enabling the potential for high coulombic efficiency plating and stripping cycles. Theoretical average Zn electrodeposition thickness (h) on wire surface can be calculated as:

$$\begin{eqnarray}&&h=\sqrt{\frac{{V}_{{Zn}}}{\pi {L}_{{wire}}}+{r}^{2}}-r\end{eqnarray} \tag{ 2 }$$

where r is the Cu wire radius, V_Zn is the deposited Zn volume based on deposition capacity, and L_wire is the effective length of wires in the wire mesh anode. The Zn plating deposition thickness is affected by many factors including plating capacity, porosity, flow condition and applied current density. For comparison, uniform distribution of Zn plating over the entire mesh surface area would produce an 11 μm thick coating for capacity of 19 mAh cm⁻², while flow produces a 32 μm average thickness in 0.5 M ZnSO₄ electrolyte at Pe = 10 (0.0152 mm s⁻¹). Thus, the Zn plating is nonuniform and/or porous. Compared with stagnant electrolytes, the flow increases the compactness of Zn plating and uniformity of the deposition thickness by decreasing the ratio of the average deposition to the theoretical thickness, improving the potential for high cycle life.

Zoom In Zoom Out Reset image size

Potentiostatic tests are performed on a Zn strip and a Cu wire mesh separated by a 2 mm gap using 0.01 and 0.02 M electrolytes held at constant 1.2 V for 1 min while the current density is recorded. These tests investigate the influence of flow on deposition morphology around overlimiting conditions. We choose dilute solutions because tests of higher molarity electrolytes produce too much gas at the voltages needed for overlimiting conditions. The Sand's time (time of minimum current) is:

$$\begin{eqnarray}&&{{t}}_{{s}}=\frac{{\text{\unicode{x003C0}D}}{\left({zFc}\right)}^{{\bf{2}}}}{{\bf{4}}{\left({j}\left({\bf{1}}-{{t}}_{+}\right)\right)}^{{\bf{2}}}}\end{eqnarray} \tag{ 3 }$$

where z is the electron number, F is the Faraday constant, c is the concentration of Zn²⁺, j is the current density, and t₊ is the cation transference number. ³⁷ The Sand's time is the characteristic time that indicates the salt concentration at the electrode surface approaches zero. ⁴¹ Figure 5 shows that the current density initially grows and then drops to a minimum, indicating diffusion-limited conditions associated with ion concentration reaching zero at the anode surface. The increased current after reaching the minimum possibly results from the electroconvective flux to the electrode surface due to unstable convective fluid motions driven by the electric field and the initiation of gassing. Using the diffusivity values above, the calculated Sand's time of 0.01 M and 0.02 M are 18 s and 11 s, respectively, comparable to the experiential values of 26 s and 19 s. In 0.01 M electrolyte, a small flow rate (Pe = 0.4) slightly changes the shape of the current response and increases minimum current density from 2.3 mA cm⁻² without flow to 2.9 mA cm⁻². For ${Pe}{\boldsymbol{\geqslant }}{\boldsymbol{1}},$ the current density increases monotonically, so overlimiting conditions are not observed. The current density also increases with flow rates. Similar phenomena are observed for 0.02 M electrolyte in Fig. 5b. Figure 5c shows SEM photos of the Zn deposition at 1.2 V after 10 min in 0.01 M stagnant electrolyte. The dendritic structures are more pronounced in the stagnant dilute solution than in the previous results for moderate (0.5 M) and concentrated (1 M) solutions. Dendrites are suppressed by electrolyte flow at Pe = 2 (Fig. 5d). Similar morphological improvements for 0.02 M electrolyte are observed in Figs. 5e, 5f.

Zoom In Zoom Out Reset image size

In Fig. 6, cycling tests are performed on a Zn-Cu asymmetric cell consisting of a zinc strip and copper wire mesh placed 2 mm apart in 0.5 M ZnSO₄ electrolyte at a current density of 50 mA cm⁻² (${\boldsymbol{0}}{\boldsymbol{.}}{\boldsymbol{65}}{{i}}_{{RS}}$). The copper wire mesh is plated for 3 min and stripped to a cutoff voltage of −0.5 V for four cycles, followed by a full strip to cutoff voltage of −0.7 V. Figure 6 shows that in stagnant electrolyte, the cell exhibits higher polarization and short circuits after 18 cycles. At Pe = 33, however, the cell cycles for over 1300 cycles at a Coulombic Efficiency (CE) of over 80% in the first 400 cycles. Electrochemical Impedance Spectroscopy (EIS) tests in Fig. s4 show that impedance is improved by increasing flow velocity in 0.5 M ZnSO₄ electrolyte.

Zoom In Zoom Out Reset image size

A two-dimensional (2D) model in COMSOL Multiphysics (COMSOL Inc.) provides the flow field, concentration and current density distributions at various flow rates. A detailed investigation of the influence on metal electrodeposition by flow-through and flow-by electrolytes have been reported by Parekh et al. ^26–28,42 We choose a single wire to represent the wire mesh due to its repetition and symmetry. Figure 7a shows the simulation geometry and flow field at Pe = 1. Normal inflow velocities are applied to the inlet, and no pressure boundary condition is set on the outlet. Symmetry boundary conditions on the left and right model the periodic array of wires in the mesh. The 30 μm diameter half circle represents the half cross section of a wire in the mesh electrode, and this surface is assumed to have uniform potential and to sink the entire current with the surface distribution calculated by COMSOL. The flow velocity at the wire surface is zero determined by no-slip boundary condition. The streamlines converge and flow rate accelerates between wires in the mesh. Modules of laminar flow, transport of dilute species and secondary current distribution are coupled to elucidate the flow and ion transport throughout the 2D domain. The model parameters are shown in Table II. The governing equations and boundary conditions are listed in supplementary materials.

Zoom In Zoom Out Reset image size

Figures 7b–7d show how 0.01 M flowing electrolyte affects the ion concentration distribution under a current density of 1.2 mA cm⁻² applied uniformly across the inlet. The average current density normalized by the wire surface is approximately 1.8 mA cm⁻². In the stagnant electrolyte shown in Fig. 7b, the concentration is almost zero (0.3 mol m⁻³) at the bottom of the wire, close to limiting conditions. Flow-through electrolyte (Pe = 0.1) decreases the concentration gradient with a minimum of 1.91 mol m⁻³ at the bottom of the wire. At 0.0246 mm s⁻¹ (Pe = 1), the concentration distribution is more uniform with 7 mol m⁻³ minimum concentration across the domain. The uniformity of concentration at Pe = 1 shows how flow mitigates concentration polarization by replenishing zinc ions to the electrode surface.

Figure 8 shows the current distribution around the wire. The origin of the arc length coordinate is the bottom of the wire, and the arc length increases in the clockwise direction. In Fig. 8a, the maximum and minimum current densities are (2.0 and 1.6 mA cm⁻²) and (1.9 and 1.8 mA cm⁻²) at Pe = 0.1 and Pe = 1, respectively. In contrast, the local current density in stagnant electrolyte ranges from 0.8 to 2.4 mA cm⁻² due to uneven ion concentration distribution. The extremums are located at the top and the bottom of the wire, consistent with the experimentally measured zinc thickness distribution for 0.5 M electrolyte (Fig. 4c).

Zoom In Zoom Out Reset image size

A needle-shaped dendritic structure is included at the top of the wire for the results in Fig. 8b. With a 10 μm dendrite (1/3 of the wire diameter), the concentration variation around the wire increases the local current density difference (Fig. 8b). The current density at the dendrite tip (3.0 mA cm⁻²) is approximately 3 times the current density at the wire bottom (1.1 mA cm⁻²), indicating preferential plating on dendrite tips. For Pe = 1, dendrite tip's current density (2 mA cm⁻²) is only slightly larger than the minimum (1.8 mA cm⁻²) at the wire bottom, indicating the mitigation of dendrite growth. This illustrates that flow-through electrolyte can suppress dendrite growth by maintaining a homogeneous current distribution.