Effect of Concentration of Tetraethylammonium Tetrafluoroborate on the Cell Performance of Electrochemical Double-layer Capacitors

Article information

Abstract

INTRODUCTION

Tetraethylammonium tetrafluoroborate (TEA BF₄) is a salt widely used in acetonitrile (AN)-based electrolytes for electrochemical double-layer capacitors (EDLCs) because of its low cost and reliable performance. Although TEA BF₄/AN electrolytes have been extensively used in various EDLC applications, such as rail vehicles [1], renewable energy [2], low-temperature energy storage [3], and hybrid vehicles [4], the growing demand for high-energy and high-power EDLCs necessitates the development of new electrolyte materials. Commercial EDLCs utilizing TEA BF₄/AN typically supply 2.7 V, whereas newer products operate at up to 3.0 V. Increasing the maximum voltage of the capacitor is the most effective strategy for meeting these performance demands. While the presence of impurities and functional groups on activated carbon [4–8] can influence the cell voltage, the electrochemical potential window (EPW) formed during electrolysis [9–12] is the dominant factor limiting the maximum cell voltage. TEA BF₄/AN cells are operated at up to 2.7 V at 25°C without significant deterioration because of the electrode passivation layers [10,13,14] that suppress electrolysis. However, these passivation layers weaken with increasing temperature, resulting in an accelerated electrolysis rate [15,16]. Therefore, the maximum voltage should be reduced to maintain cell lifetime at high-temperature conditions.

In addition to the maximum voltage, current is another critical factor that influences EDLC performance, which is generally depicted in the Ragone plot. TEA BF₄ exhibits good solubility in AN and is less viscous than several other battery solvents [17,18], making TEA BF₄/AN one of the preferred electrolytes for non-aqueous EDLCs. However, no criteria or rationale have been established to select the optimal concentration. Depending on the power requirement, the highest concentration is advantageous for maintaining low stress on the cells, resulting in a long cycle life. If the electrolyte salt concentration is too low, the electrode may not deliver its maximum capacitance due to insufficient ion availability. Conversely, if the concentration is too high, other ions may hinder ion mobility, and the increased viscosity may have an adverse effect.

This study aimed to provide a comprehensive understanding of the effects of concentration on cell performance. Electrolyte concentration affects every aspect of cell behavior, including power, energy, and cycle life. Changes in the electrolyte salt concentration alter the solvent concentration in the electrolyte solution. The findings of this study will aid in optimizing electrolyte compositions for specific applications and identifying strategies to overcome current limitations for potential improvements.

MATERIALS AND METHODS

RESULTS AND DISCUSSION

An electrolyte concentration of 1 M is typically selected for most electrochemical cells, including those used in many EDLC applications. Choosing an adequate concentration for specific applications can improve cell performance and reduce costs. Solubility tests of TEA BF₄ in AN were conducted to determine its maximum usable concentration. Visual analysis indicates a solubility limit of approximately 1.65 M at 25°C. As the solubility of the salt differed from its ionization concentration, the maximum concentration was not used in the cell tests. Instead, solutions ranging from 0.3 to 1.5 M were used to investigate concentration dependence on the cell performance. Fig. 1 shows the variations in viscosity and conductivity as a function of concentration. Both viscosity (Fig. 1A) and conductivity (Fig. 1B) deviated from linearity when the concentration approached 0.9 M. The deviation in the conductivity was significantly higher than that in the viscosity. The change in conductivity between 1.2 and 1.5 M is significantly small owing to the rapid increase in viscosity and limited ionization of TEA BF₄ in AN above 1.2 M. A conductivity comparison between 25 and 50°C is shown in Fig. S2 to observe the changes in ionic movement due to temperature. The increase in temperature enhances conductivity more significantly at 1.2 and 1.5 M than at lower concentrations. An increase in temperature reduces the viscosity and enhances the solubility of salts, thereby facilitating ionic activity more effectively at high concentrations.

Fig. 1.

Conductivity and viscosity as a function of concentration. Conductivity and viscosity were obtained at 25±1 and 25±2°C, respectively. (A) Viscosity vs. concentration (B), conductivity vs. concentration.

Fig. 2A shows the rate capabilities at various electrolyte concentrations. The specific capacitance values shown in Fig. 2A are half-cell values. The lowest concentration (0.3 M) is insufficient to supply sufficient ions to achieve the maximum capacitance of the cell. The number of ions in the low-concentration electrolyte was significantly lower than the number of available adsorption sites on the electrodes. At 0.6 M, sufficient ions are available to saturate the adsorption surface at low current densities. However, the number of ions at 0.6 M is insufficient to respond to high current densities, such as 5 A g^–1 and higher current densities. As the current density increased, the low-concentration electrolytes exhibited a reduced ability to supply ions in response to the rising current. Consequently, the lower-concentration electrolytes demonstrated relatively lower capacitances as the current increased. At the highest current density, 20 A g^–1, the capacitance was found to align with the order of the electrolyte concentrations.

Fig. 2.

Rate capability tests with electrolyte concentra are 0.1, 0.5, 2, 5, 7, 10, 15, and 20 A g^–1. The plotted v and discharge processes were performed at identica cells. (A) Rate capability tests (B) Ragone plots

The rate capabilities shown in Fig. 2A were converted into Ragone plots (Fig. 2B). The energy density of the full cell and power density were calculated using Eqs. (3) and (4) [20,21].

where V is the maximum cell voltage. m_c is the mass of AC in the cell (kg), and t is the discharge duration (h). The Ragone plot illustrates the energy density at the maximum power in the following order: 1.5 M > 1.2 M > 0.9 M > 0.6 M > 0.3 M. As the current density increased, higher concentrations were advantageous for increasing the energy density. However, if the AC contains more micropores, the order may change because the increased viscosity at high concentrations can slow ion movement within narrow pores [22]. The same cells were run to obtain the rate capability and Ragone plot at 50°C shown in Fig. S3. At low current densities, the capacitances shown in Fig. S3A displayed random variations less dependent on the electrolyte concentration. High-concentration electrolytes, known for their superior energy density at 25°C, consistently demonstrated better performance as the current density increased. However, at the highest current density of 20 A g^–1, electrolyte performance differences became less pronounced. This effect can be partially attributed to the increased self-discharge rate at higher temperatures than at lower temperatures [23].

At low currents, the movement of ions can be influenced by the surrounding ions and viscosity. Consequently, the correlation between electrolyte concentration and capacitance is relatively weak under low-current conditions. However, a higher electrolyte concentration is advantageous for effectively supplying ions to satisfy the increasing current demand under high-current conditions. Therefore, from the perspective of capacitors as energy storage devices providing high power output, higher electrolyte concentrations can improve performance, as shown in Figs. 2B and S3B. Additional temperature dependence was observed with new cells at lower temperatures of 5 and 15°C, as shown in Fig. S4. Low temperatures slow ion movement, which directly decreases the capacitance values. High concentrations, 1.2 and 1.5 M, lose significantly more capacitance than all other concentrations. The viscosity of electrolytes strongly influences capacitance at lower temperatures. The increased viscosity of the electrolytes caused by lowering the temperature subsequently decreased the capacitance more at high concentrations than lower concentrations. The order of the capacitance among electrolytes is: 0.6 M  0.9 M > 0.3 M  1.2 M > 1.5 M. 0.3 M, which has the lowest viscosity, however, supplies insufficient ions for the available adsorption surface. Although coin cells are more strongly influenced by resistance than large cells, the rapid capacitance decreases at 15°C (Fig. S4B) are substantial. The capacitance at 5°C and 20 A g^–1 renders cells with a 1.5 M concentration nearly unusable (Fig. S4A). Low temperatures increase viscosity and decrease solubility, making ion movements more sluggish and decreasing the number of usable ions. Electrolytes with concentrations close to the solubility of salt may have limited use at low temperatures, as their performance deteriorates rapidly at low temperatures.

The time constants (τ) of the capacitors were measured using chronoamperometry. The capacitive current decay during the potential step was determined using Eq. (5):

where E is the magnitude of the potential step, R_s is the electrolyte resistance, and C_d is the cell capacitance. The initial current during the potential step decreased exponentially. The time constant (R_sC_d) is the time required for the current to decay to 37% of its initial value. Although capacitors exhibit very low faradaic currents near the cutoff voltage, the measured currents include both faradaic and capacitive currents. When the faradaic current is non-negligible, Eq. (5) could not be used to determine the time constant. Furthermore, the current vs. time plot became inconsistent when a Faradaic current was present. To address this, the cells used in the rate capability tests were preconditioned at a holding voltage of 2.7 V for 5 h, reinforcing the surface protection layer and reducing the faradaic current to negligible levels.

The cells with varying salt concentrations were stepped to 2.7 V to observe the change in current as a function of time (Fig. 3). The time constants of the cells indicate their relative responsiveness. The cellular response rate increased with increasing concentration. Because the time constant is the product of electrolyte resistance and cell capacitance, an electrolyte with higher conductivity allows the cell to operate faster. A lower electrolyte concentration results in a lower capacitance. However, as the concentration decreased, the increase in resistance became more significant, making the electrolyte resistance a major factor influencing the time constant. The time constants in Table 1 demonstrate an efficient supply of ions for adsorption at high concentrations. This explains the observed order of the energy density at the maximum power across the electrolytes. However, cells containing high-energy-density ACs with more micropores may exhibit a different relationship between the conductivity and time constant. Bulk ion conductivity does not directly govern the ion movement within narrow pores.

Fig. 3.

Chronoamperometry of the cells at a stepped voltage of 2.7 V. To minimize the faradaic current, the voltage was maintained at 2.7 V for 5 h before measurement.

Table 1.

Time constant (τ) of cells at different concentrations of TEA BF₄.

Cell responsiveness is primarily affected by the electrolyte conductivity. To operate cells at a high-power output for a long time, the electrochemical durability of the material is crucial. Therefore, the relationship between the durability and salt concentration was investigated. An increase in the salt concentration reduced the proportion of solvent. Electrolytes with varying concentrations can help identify which of the two components, solvent or salt, has a greater impact on the lifespan of the electrolyte. Accelerated durability tests were conducted using the electrolytes to identify the main factors limiting the lifespan of the cells. Cells were charged to 3.0 V (instead of 2.7 V) and held at 70°C. Specific capacitance retention was recorded every 50 h at 25°C with a 2.7 V cutoff. To compare the capacitance retention with the initial values at 25°C, the capacitance of the cells was measured after the cells were cooled at 25°C for 24 h. The capacitance loss due to the stimulation of voltage and temperature is shown in Fig. 4A, plotted using the data from Fig. S5. Cells with concentrations of 0.3 M and 0.6 M exhibit a rapid loss of capacitance over time, whereas cells with concentrations above 0.9 M show significantly greater durability compared to those with lower concentrations. Increasing concentration improves the cell lifetime. Although the temperature stimulation slightly recovered the cell capacitance after 200 h, the capacitance continued to decline. Cell deterioration was determined by monitoring cell resistance over time. The cell resistance values depicted in Fig. 4B were determined from the IR drop observed during the capacitance measurements, as presented in Fig. S5. Cell deterioration was suppressed with increasing concentrations. The results shown in Fig. 4 indicate that the degradation of the cells during cycling and temperature stress was due to solvent degradation rather than the salt. While the degradability of TEA BF₄ cannot be entirely ruled out because the cation structure improves durability [24–26], the durability tests in this experiment indicate that the solvent is more prone to degradation than TEA BF₄.

Fig. 4.

Accelerated cell durability tests at various electrolyte concentrations: (A) specific capacitance retention over time and (B) cell resistance over time.

An increase in cell resistance and loss of capacitance was observed during the visual examination of the electrodes following the durability test. A cell that exhibited a significant capacitance loss during this test was disassembled for postmortem analysis, wherein the electrodes were carefully inspected. During this process, the separator between the electrodes is removed, causing damage to certain areas of the electrode. Despite minor damage to the electrodes, a noticeable difference was observed in the physical properties of the two types. The positive electrodes are more rigid and brittle than the negative electrodes. Fig. 5A shows the comparative images of the two electrodes after bending. The positive electrode crumbled like a thin piece of chalk and easily detached from the current collector, whereas the negative electrode was soft enough to adhere to the current collector and remained intact when bent. Regardless of concentration, all the capacitors exhibited similar characteristics, albeit to varying degrees. Under these experiment’s conditions, the positive electrode’s stiffness was identified as a key factor leading to a significant decrease in cell lifespan by increasing the contact resistance between the electrode and the collector. SEM images of the electrodes after the durability test are presented in Fig. 5B, showing the electrodes in an unbent state. Cracks in the electrodes occurred solely due to the accelerated durability testing without any physical impact. This suggests that the functionality of the binder in the positive electrode deteriorates far more than that in the negative electrode as the lifespan of the capacitor decreases. Our binder was composed of both CMC and SBR, which are used in many capacitor manufacturers. The electrochemical and physical properties of CMC and SBR composites should be improved to broaden their operational voltage and temperature ranges. A new binder material, potato starch (PS) [27], was proposed to enhance the physical durability of the electrode by adding flexibility to a CMC-containing electrode. However, a more flexible binder made from poly(tetrafluoroethylene) (PTFE) induced a crack in the positive electrode above 3.0 V, regardless of the temperature increase [28]. It implies that all commercial binders crack when the cell voltage increases. Temperature can expedite the formation of cracks; however, the primary cause is an increase in cell voltage, which subsequently increases the potential of the positive electrode. Gas generation and exfoliation of the graphene layer have been reported as a direct cause of crack formation. Corrosion of the current collector [29] is a possible reason for the electrode detachment. The stiffness of the electrode and the corrosion of the aluminum cause the electrode to detach from the collector. The gas formation and exfoliation of the graphene layer, accelerated by the temperature, can damage the positive electrode, as shown in Fig. 5. AC destruction of the positive electrode at 3.5 V was previously reported [30]. Although the cell voltage in our study is 2.7 and 3.0 V, somewhat lower than those in the reported studies, accelerated aging due to temperature can still cause structural modifications to AC.

Fig. 5.

Images of the negative and positive electrodes after the accelerated durability testing of the cell. (A) Photographs and (B) SEM images

The width of the EPW of the electrolyte was sufficiently high to accommodate a 2.7 V cell voltage. However, the voltage region of the capacitor is skewed toward the boundary of the positive voltage in ESW. Therefore, the electrolysis intensity is higher at the positive electrode [10,14,15,30] than at the negative electrode. The passivation film effectively suppresses electrolysis during mild cell operation. Although the electrolysis intensity was low, the consumption of the electrolyte and the accumulation of decomposed products over time gradually increased the cell resistance and reduced the capacitance, as shown in Fig. 4. A recent publication reported the formation of a passivation layer composed of decomposed electrolyte [31]. The degradation intensity increases with increasing temperature and cell voltage. Monitoring the electrolysis current as a function of concentration is another supportive measure for identifying easily decomposable materials in electrolytes. The electrolysis intensity was evaluated before the formation of the passivation film. Therefore, LSVs for electrolysis should be conducted using fresh electrodes to avoid interference from passivation films. The EPWs are presented in Fig. S6, where the oxidation intensity was significantly higher than the reduction intensity. Since −1.6 or −1.7 V represented the negative limit for a 2.7 V cell voltage [19], electrolysis during the negative scan is not discussed in this context. Fig. 6 shows only the positive-scan LSVs. As the positive potential increases in the LSV, the intensity of electrolysis increases sharply, causing a noisy current due to interference from vigorous bulk electrolysis activities[15]. The vertical dashed line at 1.0 V in Fig. 6 indicates the positive limit for cells for a 2.7 V cell voltage, which is used to compare electrolysis currents. Higher electrolysis currents were observed at lower concentrations, indicating that the solvent contributes more to the electrolysis current than the salt. This explains the rapid increase in resistance and poor cell durability at low concentrations.

Fig. 6.

Electrolysis current on the positive electrode. The scan was initiated at an open-circuit potential with a scan rate of 10 mV s^–1. The working and counter electrodes in the cells were AC electrodes.

The results in Figs. 4–6 suggest that the solvent is the primary component of the electrolyte responsible for the rapid degradation due to electrochemical oxidation [32–34]. As electrolysis becomes stronger at lower concentrations, gas evolution [35] and exfoliation of the graphene layer become more intense, leading to a shorter cycle life at lower concentrations. Curves in Fig. 6 show that AN is easier to decompose than salt. In addition, the binder and AC, solid elements in the electrode, were sufficiently stable to show smooth ESW curves and cut-off voltages at room temperature; otherwise, an unstable current would be observed. The positive electrode becomes dry, as more solvent is consumed on the positive electrode at a higher voltage and/or higher temperature. The dry electrode is physically susceptible to cracking owing to the physical impact caused by gas evolution. The oxidation of AN produces a decomposed product on the positive electrode, which blocks the pores of the AC and increases the cell resistance [30]. Oxidation of AC and the composition of carbon altered by oxidation. However, the oxidation onset potential varies depending on operating conditions. Intense electrolysis also alters the physical properties of the AC [36,37] in the positive electrode. Fig. S7 is a schematic of cell deterioration through cycles, where cell components are gradually damaged, resulting in cell failure.

Increasing concentration from 1 M to 1.5 M extends the cycle life and improves the energy density. The extent of performance improvement varies with the operating temperature; however, high-concentration electrolytes consistently contribute to enhanced performance. AN-based electrolytes, which have low viscosity [38] and exhibit good conductivity, offer high-power output; however, the strong electrolysis of AN at the positive electrode limits the improvement in cell voltage. The proactive incorporation of passivation layers on the electrodes can compensate for the narrow EPW. Precise control and meticulous efforts are required to form these layers, especially in capacitors, where ion adsorption and desorption are critical.

CONCLUSIONS

Understanding the material properties is essential before modifying them for use in EDLCs. Investigating the effects of concentration on cell behavior must be prioritized before exploring new electrolytes. Highly concentrated electrolytes allow for fast cell operation. The RC time constants demonstrate that higher concentrations benefit rapid cell operation. Concentrations closer to the solubility limit of the salt were more favorable for increasing the energy density at the highest power densities at temperatures from 25 to 50°C. Electrolytes with concentrations close to their solubility may sometimes have limited use at lower temperatures than 25°C, as their performance deteriorates rapidly at low temperatures. Accelerated durability tests revealed that cells with higher TEA BF₄ concentration exhibited longer cycle lives, underscoring the electrochemical stability of AN in the electrolytes. After disassembling the capacitor used in the durability test, the positive electrode became brittle and detached easily from the current collector. The physical properties of the positive electrodes were severely degraded. Therefore, determining new solvents and binders resistant to oxidation is necessary to increase cell voltage and cycle life.

SUPPORTING INFORMATION

Article information Continued