Abstract
Electrodeposition of zinc at current densities close to the mass transport limit produces needle-like dendrites. Suppressing dendrites is of technological interest to applications in sacrificial corrosion protection coatings and flow batteries. In the present work, we report the use of polyethylene glycol (PEG, M.W. = 200) as an effective electrolyte additive to suppress dendrites during zinc electrodeposition from halide-based electrolytes. Dendrite growth rate is measured as a function of the PEG concentration using in situ optical microscopy, which shows that the dendrite suppression efficacy due to PEG increases with PEG concentration. Polarization experiments on a rotating disk electrode provide system parameters, i.e., the exchange current density and the cathodic transfer coefficient, which confirm that PEG suppresses the zinc electrodeposition kinetics. The kinetic parameters are incorporated into a simple electrochemical model for activation-controlled dendrite propagation. The model predicts an order of magnitude reduction in the zinc dendrite growth rates in the presence of high concentration of PEG (10000 ppm), consistent with experimental findings.
Export citation and abstract BibTeX RIS
Electrodeposition of zinc (Zn) is a technologically important electrochemical process. Due to its electronegativity (EZn = – 0.76 V vs. SHE), electroplated Zn is used widely in sacrificial corrosion resistant coatings.1 Electrowinning of Zn involves its electrodeposition from acidic ZnSO4 electrolytes.2 Zn electrodeposition is gaining interest in the context of advanced large-scale energy storage devices such as flow batteries. Specifically, Zn electrodeposition from zinc halides is central to the charging of Zn-halogen (Zn-bromine or Zn-chlorine) flow batteries.3
In commercial applications, a high rate (current density) of Zn electrodeposition is almost always desired. This is necessitated by throughput, cost, and performance considerations. However, as the Zn electrodeposition rate approaches the transport-limited rate (i.e., the limiting current density), the electrodeposits obtained are rough, powdery, and dendritic. This deposit morphology is detrimental to process performance. In a Zn-based flow battery, dendrites degrade battery cyclability, round-trip efficiency, and life.4
A popular technique for controlling surface morphology in electrodeposition is the use of ppm levels of plating additives. Suppressing, leveling, brightening, and surface film-forming additives have been widely employed.5–7 In the context of Zn, numerous investigators have studied the effect of additives such as benzoic acid, quaternary ammonium compounds, and Triton X-100.8–15 However, these studies have been primarily aimed at understanding the effect of the aforementioned additives on the electrodeposit crystal structure.10,13–15 No quantitative correlations relating the effects of additives on the Zn dendrite growth rates were provided. In recent studies, the synergistic roles of additives such as benzoic acid, polymeric ethers, and amines in controlling surface roughness and hydrogen co-evolution during Zn electrodeposition were investigated.8,9,16 While these studies provided valuable insight into the additives' adsorption mechanisms, no direct quantification of the dendrite suppression efficacy of these additives was elucidated.
In the present paper, we focus on the use of polyethylene glycol (PEG, M.W. = 200) additive to suppress Zn dendrites formed during electrodeposition from halide electrolytes. This choice of PEG M.W. = 200 is based on its stability and resistance to breakdown or degradation during the prolonged operation of a Zn battery. Using in situ optical microscopy, we monitored the growth of Zn dendrites, which enabled quantitative characterization of the dendrite suppression efficacy of PEG in various halide electrolytes. Electrode polarization measurements during Zn electrodeposition on a rotating disk electrode in the presence of PEG elucidated the role of PEG in suppressing electrodeposition kinetics. These kinetic parameters, when incorporated into a simple electrochemical model of activation-controlled dendrite growth, provide dendrite suppression efficacies that are in reasonable agreement with experiments, thereby providing a predictive tool for characterizing the effect of additives in Zn dendrite suppression.
Experimental
Materials
Electrodeposition of Zn was studied from aqueous electrolytes containing zinc chloride (ZnCl2) or zinc bromide (ZnBr2). Both salts were purchased from ACROS and used as received. The additive polyethylene glycol (PEG, average M.W. = 200 g/mol) was acquired from Sigma-Aldrich and was also used as received. Millipore-spec (18.2 MΩ) deionized water was used to prepare all aqueous electrolytes. Prior to preparing the electrolytes, high purity argon was vigorously bubbled through the deionized water for one hour to remove dissolved oxygen. Electrolyte conductivity was measured using an Oakton PC2700 meter with an accompanying conductivity probe.
Methods
The setup used to track dendrite growth during Zn plating consisted of an electrochemical cell with a three-electrode configuration. A shallow glass dish contained the electrolyte, which was either 0.1 M ZnCl2 or 0.1 M ZnBr2 with varying levels (0, 100, 1000 or 10000 ppm) of PEG additive. The working electrode was the exposed tip of a No. 18 AWG PVC-insulated copper wire. Prior to each experiment, the exposed copper surface of the wire was first polished with a 600 grit sand paper and then plated for 30 minutes at 10 mA/cm2 to form a thin (∼8 μm) and smooth layer of Zn. A platinum wire (Encompass) was used as the counter electrode and a Ag/AgCl electrode filled with 3 M KCl saturated with AgCl (Radiometer Analytical) served as the reference electrode. The tip of the wire electrode at which Zn was electrodeposited was monitored during Zn plating by a National 410 stereo optical microscope placed directly above the horizontally aligned wire electrode. The optical microscope was attached to a digital camera for image recording. The three electrodes were connected to a Solartron model 1287A potentiostat with automated data acquisition.
Electrode polarization during Zn electrodeposition in the presence of PEG was characterized on a rotating disk electrode (Pine). The platinum surface of the disk electrode was plated with a thin (∼3 μm) layer of Zn prior to each measurement. Rotation speed of the disk electrode was maintained at 300 rpm.
Results and Discussion
In this section, we first report the suppression of Zn dendrites by PEG additive during Zn electrodeposition from aqueous electrolytes containing ZnCl2 or ZnBr2. The effect of PEG concentration on dendrite suppression efficacy is also reported.
Dendrite suppression in a ZnCl2 electrolyte
The aforementioned in situ optical microscopy setup was employed for characterizing Zn dendrite suppression by PEG. The working electrode (Zn-coated wire) was held at a constant potential and the growth of dendrites was tracked as a function of time. Two electrode potentials (Vapp) were tested: –1.25 V and –1.30 V (vs. Ag/AgCl).
After 8 minutes of plating at Vapp = –1.25 V, microscopy revealed rapid development of dendritic morphology at the wire electrode in the absence of PEG (panel (a) of Figure 1). Needle-like dendrites several hundred microns in length were visible. At Vapp = –1.30 V as well, dendritic electrodeposition was evident (panel (e) of Figure 1). Dendrites grew faster at –1.30 V as compared to –1.25 V.
Addition of 100 ppm of PEG to the electrolyte suppressed dendrite formation marginally (panels (b) and (f) of Figure 1). Fractal-like aggregates were still visible at both electrode potentials indicating that 100 ppm of PEG additive is insufficient for eliminating dendritic morphology. However, marked suppression of dendrite growth could be obtained at PEG concentrations of 1000 and 10000 ppm. At these PEG concentrations, deposits did not exhibit the classical dendritic morphology.17 On the contrary, deposits were more compact, indicating effective suppression of dendritic morphology. At 10000 ppm of PEG, dendrites were completely eliminated at –1.25 V and significantly suppressed at –1.30 V.
To ensure that the addition of PEG to the ZnCl2 electrolyte does not accelerate side reactions (such as hydrogen co-evolution) that may degrade the current efficiency of Zn electrodeposition, we electrodeposited Zn galvanostatically (at 5 mA/cm2) from the aforementioned electrolyte onto a Zn-seeded substrate in the presence and absence of PEG. Electrodeposit weight gain was measured and converted into current efficiency by applying Faraday's law. The current efficiency in all tested cases (in the presence and absence of PEG) exceeded 95% indicating no detrimental effects of PEG on the Zn deposition current efficiency. This confirms that PEG suppresses Zn dendrites while maintaining an overall high efficiency of Zn plating.
Figure 2 shows chronoamperometry plots corresponding to the dendrite growth studies reported in Figure 1. Chronoamperometry plots are shown only for the applied potential of –1.25 V since the general trends were similar at the applied potential of –1.30 V. Upon application of the potential, the Zn deposition current decayed from about 50 mA/cm2 (at time ∼0 s) to about 29 mA/cm2 over an 8 s period. This decay is attributed to the development of the concentration profile in the Nernst diffusion boundary layer near the electrode surface. The steady state current of 29 mA/cm2 was relatively insensitive to the applied potential, thus representing deposition of Zn close to the system limiting current density. This was also confirmed by polarization scans on the wire electrode, which showed that the Zn limiting current density at the wire electrode was ∼30 mA/cm2. As seen in Figure 2, after the first transient diffusion period (∼8 s), the current remained constant for a period of about 100 s. During this period, no Zn dendrites were observed (through the microscope) on the wire electrode. However, after this period, the current gradually increased. The rate of current increase depended on the concentration of PEG in the electrolyte (Figure 2). The most rapid increase in current was recorded in the absence of PEG. Dendritic morphology concurrently developed on the wire electrode, which suggests that the current increase is attributed to the surface area increase during dendritic deposition. The rate of current increase (and thus surface area increase) was suppressed in the presence of PEG with 100 ppm and 1000 ppm PEG showing lower rate of rise. At 10000 ppm PEG, the current did not increase measurably, remaining nearly constant for about 8 min. This confirms the near complete suppression of morphology evolution due to Zn dendrites. These electrochemical trends complement the in situ growth studies (Figure 1) in validating the concentration-dependence of the Zn dendrite suppression by PEG additive.
Dendrite suppression in a ZnBr2 electrolyte
In zinc-bromine flow batteries, aqueous solutions of ZnBr2 are employed as electrolytes.3 To demonstrate that the PEG additive functions as an effective dendrite suppressor during Zn deposition from a ZnBr2 system, we conducted experiments using the same in situ microscopy setup employed above for studies with ZnCl2, but replacing the ZnCl2 electrolyte with a 0.1 M ZnBr2 electrolyte. A sample data set (collected at Vapp = –1.25 V vs. Ag/AgCl) is shown in Figure 3. Panel (a) of Figure 3 shows dendritic morphology evolution during Zn deposition in the absence of PEG. Panel (b) of Figure 3 confirms the complete suppression and elimination of dendrites in the presence of 10000 ppm of PEG in the ZnBr2 electrolyte. This demonstrates the flexibility of applying PEG as an effective dendrite suppressor in bromine-based electrolytes.
Model of Zn dendrite suppression by PEG
In this section, we propose a model for the suppression of Zn dendrites by PEG. Model parameters are determined through electrochemical polarization measurements on a rotating disk electrode. Dendrite suppression efficacy of PEG predicted by the model is compared with experimental observations reported in the previous sub-section.
Dendrite growth during metal electrodeposition has been modeled by numerous investigators. The theory of activation-controlled dendrite propagation was introduced by Barton and Bockris for silver electrodeposition18 and by Diggle et al. for zinc electrodeposition from an alkaline zincate solution.19 This framework was later extended to other systems such as zinc deposition from acidic halide electrolytes,20,21 copper electrodeposition,22,23 and lithium electrodeposition.24,25 These dendrite growth models typically analyze the various overpotentials at the dendrite tip and correlate them to the tip current density (or growth rate).19 The overpotentials include the activation overpotential, concentration overpotential, ohmic overpotential, and a surface overpotential due to the radius of curvature of a dendrite tip. At an applied potential, Vapp, a voltage balance at the dendrite tip (as shown in Figure 4) provides:
In Eq. 1, E is the reduction potential of Zn (measured to be –1.0 V vs. Ag/AgCl), ηa is the activation overpotential, ηΩ is the ohmic overpotential, ηc is the concentration overpotential and ηs is the surface overpotential due to the curved dendrite tip. In analogy to the model by Diggle et al.,19 we assume that dendrites in our Zn electrodeposition system grow purely under activation control, as shown schematically in Figure 4. To validate this hypothesis, we now examine the magnitude of each overpotential relative to the total overpotential in the system. Since experiments in the present work were conducted at Vapp = –1.25 V or –1.30 V, and E = –1.0 V (vs. Ag/AgCl), the net overpotential is either 250 mV or 300 mV in magnitude, depending on the applied potential. Since the entire electrode surface is equipotential, this net overpotential exists at the tip of each growing dendrite (Figure 4). The contribution of the ohmic overpotential to the net overpotential is:
In Eq. 2, i is the dendrite tip current density, l is the characteristic length scale over which the ohmic potential drop develops and κ is the electrolyte conductivity. The dendrite tip growth rate measured from experiments is about 40 μm/min, which corresponds to a current density i ≈ 1.5 A/cm2. The characteristic length l can be taken as the dendrite tip radius in analogy to a micro-electrode.26 Typical dendrite tip sizes are 0.5 μm as reported by Diggle et al.19 With a measured electrolyte conductivity of 17.7 mS/cm, we estimate ηΩ to be a meager 4 mV and thus negligible in comparison to the net overpotential at the dendrite tip.
Let us now examine the concentration overpotential:
In Eq. 3, R is the universal gas constant (8.314 J/mol·K), T is the temperature (298 K), n is the number of electrons transferred (n = 2) during the electrodeposition reaction, F is Faraday's constant, and iL is the limiting current density at the dendrite tip. As shown by Diggle et al.19 and Oren and Landau,21 transport of cations to the dendrite tip is controlled by 'spherical' diffusion, which leads to the following expression for the limiting current density:
Taking the diffusion coefficient () to be 7.0×10−6 cm2/s,27 the transport number (t+) to be 0.41 (calculated based on ionic conductivities reported in Ref. 27), and the dendrite tip radius r to be 0.5 μm,19 we estimate the limiting current density iL to be ∼4.6 A/cm2 at Cb = 0.1 M. Inputting this value of iL in Eq. 3, we calculate the concentration overpotential (ηc) to be ∼5 mV and thus negligible with respect to the net overpotential of about 250–300 mV. Next, we calculate the surface curvature overpotential (using the relationship used by Barton and Bockris):18
In Eq. 5, γ represents the surface tension at the electrode/electrolyte interface (7.9×10−5 J/cm2 for Zn),28 and K is the molar volume of Zn (9.2 cm3/mol). From Eq. 5, the surface curvature overpotential for a Zn dendrite with r = 0.5 μm is ∼0.2 mV, and thus also negligible.
From the above discussion, it is clear that the ohmic overpotential (ηΩ ∼ 4 mV), concentration overpotential (ηc ∼ 5 mV) and the surface curvature overpotential (ηs ∼ 0.2 mV) are all negligible in comparison to the total electrode overpotential of several hundred mV. Thus, the activation overpotential (ηa) is the dominant overpotential at the dendrite tip. This confirms that the dendrite tip grows purely under activation control. Assuming Tafel kinetics, the activation overpotential, which is almost equal to the total overpotential, takes the form:
In Eq. 6, αc and i0 represent the cathodic charge transfer coefficient and the exchange current density, respectively. These kinetic constants correlate the activation overpotential at the dendrite tip (ηa) to the current density (i) at which the dendrite grows. The dendrite growth rate, , is then obtained by applying Faraday's law:
In Eq. 7, M is the molecular weight of Zn (65.38 g/mol), and ρ is its density (7.14 g/cm3). Combining Eq. 6 and Eq. 7 yields:
Eq. 8 can be used to correlate the Zn dendrite growth rate to the system kinetic constants (αc and i0). Eq. 8 also reveals that the dendrite growth rate increases exponentially with the applied potential. While kinetic parameters for Zn electrodeposition in the absence of additives have been reported,20 no systematic study on the PEG-concentration dependent Zn electrodeposition kinetics is available. Below, we report polarization measurements on a rotating disk electrode to elucidate the effect of PEG on Zn electrodeposition kinetics.
Effect of PEG on Zn electrodeposition kinetics
Tafel plots for Zn electrodeposition on a Zn-coated rotating disk electrode (at 300 rpm) from a 0.1 M ZnCl2 electrolyte are shown in Figure 5. In these Tafel plots, the activation overpotential was obtained after subtracting the IR (ohmic) and concentration overpotentials from the measured total overpotential. Newman's analytical solution for ohmic resistance of a disk electrode was used for IR correction,29 and Eq. 3 (with iL provided by the Levich equation)30 was used to correct for concentration polarization. The effect of PEG concentration on the electrode polarization is evident from Figure 5. The electrode polarization increases monotonically with PEG concentration, indicating that PEG adsorbs on the electrode surface and polarizes the electrode during Zn deposition. This is analogous to the electrode polarization observed during copper electrodeposition31,32 in the presence of PEG additive. Figure 5 shows that the slope of the ln(i) vs. ηa curve, which relates to the Tafel slope, remains unaltered in the presence of PEG. This indicates that PEG does not affect the cathodic transfer coefficient during Zn electrodeposition. From Figure 5, the cathodic transfer coefficient was computed to be: αc = 0.79. PEG substantially alters the Y-intercept of the ln(i) vs. ηa curve, indicating a substantial lowering of the exchange current density (i0) of Zn plating. The values for the exchange current density were computed from Figure 5 and are listed in Table I. The exchange current density is lowered by almost an order of magnitude by the addition of 10000 ppm of PEG to the base electrolyte. The effect of PEG on i0 (but not on αc) indicates that PEG suppresses deposition kinetics by adsorbing on the electrode surface and forming a surface passivating film. The extent of coverage of this surface film, which depends on the PEG concentration,33,34 controls the active sites available for Zn electrodeposition. The active surface area for Zn deposition manifests as a shift in the apparent exchange current density, similar to the effect of PEG on copper electrodeposition kinetics.35
Table I. Kinetic parameters for Zn electrodeposition in the presence of PEG additive, indicating that PEG predominantly affects the exchange current density during Zn electrodeposition. Base electrolyte contains 0.1 M ZnCl2. The cathodic transfer coefficient is: αc = 0.79.
PEG Concentration (ppm) | Exchange Current Density (i0, mA/cm2) |
---|---|
0 | 2.82 |
100 | 2.22 |
1000 | 1.30 |
10000 | 0.39 |
Dendrite suppression factor
To quantify the dendrite suppression efficacy of PEG, we now define a dendrite suppression factor (ξ) as the ratio of the dendrite growth rate observed in the presence of PEG to that observed in the absence of PEG:
A value of ξ well below unity suggests strong suppression of Zn dendrite growth by PEG, whereas ξ approaching unity suggests weak suppression of Zn dendrite growth. Estimates of the dendrite propagation length were obtained from microscopy images shown in Figure 1 by measuring the maximum distance from the electrode surface to which the Zn dendrites grew. These propagation lengths were then used in Eq. 9 to yield the experimentally observed dendrite suppression factor. To calculate the theoretical dendrite suppression factor, the dendrite growth rate expression (Eq. 8) was inserted into Eq. 9. Recognizing that the cathodic transfer coefficient is unaffected by PEG (Figure 5), we get the following approximation for the dendrite suppression factor:
Thus, dendrite suppression by PEG is directly related to the lowering of the exchange current density during Zn electrodeposition. Dendrite suppression factors measured experimentally (by inserting measurements from Figure 1 into Eq. 9) and those estimated from theory (by applying Eq. 10 with kinetic constants from Table I) are plotted in Figure 6 as a function of the PEG concentration. Both theory and experiment show a monotonic decrease in ξ (i.e., an increase in the dendrite suppression efficacy) with an increase in the PEG concentration. Good agreement is seen between the experimentally observed suppression factor and that predicted by theory. This demonstrates the ability of the simple model (Eq. 10) to serve as a predictive tool for determining the effectiveness of Zn dendrite suppression by PEG additive.
Conclusions
A combination of in situ optical microscopy, electrochemical polarization, and modeling is employed to study the suppression of dendrite growth during Zn electrodeposition from aqueous halide electrolytes containing PEG additive. The study leads to the following key conclusions:
- (i)Over a wide concentration range (100–10000 ppm), PEG suppresses dendrite formation during Zn electrodeposition. The degree of dendrite suppression depends on the PEG concentration; dendrite suppression is more effective at higher PEG concentrations.
- (ii)Analysis shows that a spherical micro-scale Zn dendrite tip is released from mass transport and ohmic limitations, allowing the tip to grow rapidly under activation, i.e., kinetic control. Thus, dendrite suppression can be achieved by using additives, such as PEG, which affect the exchange current density in Zn electrodeposition.
- (iii)PEG suppresses Zn dendrites by lowering the plating exchange current density. The higher the PEG concentration, the lower the exchange current density and thus, the higher the dendrite suppression efficacy.
- (iv)A simple model for activation controlled dendrite growth was developed to compute Zn dendrite suppression efficacies in the presence of PEG and compare them to experimental observations. Good agreement between model predictions and experimental observations was observed.
The experimental and modeling approach described herein provides a quantitative means of characterizing and understanding the effect of additives on dendrite formation during electrodeposition. While the present study reports on the effect of PEG additive on Zn electrodeposition, a more comprehensive study involving a wide range of additives and a broader set of electrodeposited metals and electrolyte systems is underway. This will provide a more complete understanding of the additive-induced dendrite suppression process.