Skip to main content
NCBI home page
ACS Omega logoLink to ACS Omega
. 2021 Nov 1;6(45):30800–30810. doi: 10.1021/acsomega.1c05133

Influence of Hydrotropes on the Solubilities and Diffusivities of Redox-Active Organic Compounds for Aqueous Flow Batteries

Yingchi Cheng †, Derek M Hall ‡,§, Jonathan Boualavong †, Robert J Hickey ∥, Serguei N Lvov ‡,§,∥, Christopher A Gorski †,*
PMCID: PMC8600646  PMID: 34805708

Abstract

graphic file with name ao1c05133_0008.jpg

In this study, we explored the extent to which hydrotropes can be used to increase the aqueous solubilities of redox-active compounds previously used in flow batteries. We measured how five hydrotropes influenced the solubilities of five redox-active compounds already soluble in aqueous electrolytes (≥0.5 M). The solubilities of the compounds varied as a function of hydrotrope type and concentration, with larger solubility changes observed at higher hydrotrope concentrations. 4-OH-TEMPO underwent the largest solubility increase (1.18 ± 0.04 to 1.99 ± 0.12 M) in 20 weight percent sodium xylene sulfonate. The presence of a hydrotrope in solution decreased the diffusion coefficients of 4-OH-TEMPO and 4,5-dihydroxy-1,3-benzenedisulfonate, which was likely due to the increased solution viscosity as opposed to a specific hydrotrope–solute interaction because the hydrotropes did not alter their molecules’ hydraulic radii. The standard rate constants and formal potentials of both 4-OH-TEMPO and 4,5-dihydroxy-1,3-benzenedisulfonate remained largely unchanged in the presence of a hydrotrope. The results suggest that using hydrotropes may be a feasible strategy for increasing the solubilities of redox-active compounds in aqueous flow batteries without substantially altering their electrochemical properties.

Introduction

Redox flow batteries are a promising approach for balancing temporal differences between electrical supplies and demands created by electricity production from intermittent renewable resources (e.g., solar and wind).1−4 Redox flow batteries store and produce energy using electrolytes containing soluble redox-active compounds that are reversibly oxidized and reduced at electrode surfaces.5,6 The main goal in developing redox flow batteries is minimizing their costs, with the U.S. Department of Energy having set a capital cost target of USD $150 per kWh by 2023.7 First-generation redox flow batteries used vanadium redox couples, which provided sufficiently high aqueous solubilities and cell voltages but were expensive due to the high price of vanadium.8 To decrease costs, researchers began using organic redox-active compounds as opposed to metal-based compounds.9−12 However, the energy densities of organic-based aqueous redox flow batteries have been limited by low solubilities of organic compounds in aqueous electrolytes.1,12−14

To date, the main approach to increase the energy densities of organic-based aqueous redox flow batteries has been to alter the functional groups on organic compounds to tune their solubilities, reduction potentials, and/or the number of electrons transferred.14−18 For example, substituting and repositioning the functional groups on quinone compound shifted its reduction potentials by up to 1.17 V19 and increased its solubility by over an order of magnitude.6,14,20 Similarly, replacing methyl groups on methyl viologen with more hydrophilic ammonium functional groups increased the number of usable electron-transfer steps from 1 to 2.18,21 Although such molecular engineering techniques show promising energy density improvements, they require organic syntheses with multiple steps, which can be time consuming and difficult to scale up.

An alternative, less studied approach to increase redox flow battery energy densities is to amend the aqueous electrolyte with a hydrotrope.15,16 Hydrotropes are amphiphilic organic compounds most commonly used to increase the solubilities of sparingly soluble organic molecules.22−28 Hydrotropes differ from surfactants in that they do not form micelles,22,26 but the mechanisms by which they promote solubilization remain a topic of debate in literature.24,26,27 Two recent flow battery studies used high concentrations of hydrotropes to successfully increase the aqueous solubilities of organic redox-active compounds. The addition of 3.0 M hydrotrope nicotinamide (vitamin B3) increased the aqueous solubility of flavin mononucleotide from 0.1 to 1.5 M in 1.0 M KOH.16 In another study, the addition of 4.0 M hydrotrope urea increased the solubility of hydroquinone (benzene-1,4-diol) from 0.5 to 1.5 M in deionized water.15 In these studies, hydrotropes did not change the formal reduction potential values of the compounds. In the case of hydroquinone, urea did not significantly alter its diffusion coefficient, and it slightly decreased its standard rate constant.15 These results indicate that hydrotropes may be an effective way to increase the solubilities of organic redox-active compounds in aqueous electrolytes when their initial solubilities are relatively low (≤0.5 M). What remains unclear is whether hydrotropes can also increase the solubilities of highly soluble redox-active compounds (>0.5 M) and, if so, how hydrotropes alter the electrochemical properties of the compounds.

The goal of this study was to identify promising combinations of hydrotropes and redox-active organic compounds that would increase the compounds’ solubilities and to determine if their electrochemical properties would change. We surveyed five redox-active organic compounds previously used in aqueous redox flow battery studies: methyl viologen (MeV),11 4-OH-TEMPO,11 anthraquinone-2,7-disulfonate (AQDS),13,29,30 4,5-dihydroxy-1,3-benzenedisulfonate (BDS),29,30 and riboflavin-5′-monophosphate (FMN)16 (Figure 1A) in the presence of five different hydrotropes: sodium xylenesulfonate (NaXS),22 sodium para-toluenesulfonate (NapTS),22 proline,22 urea,15,28 and tert-butanol (TBA)31 (Figure 1B). We selected the five organic compounds because they have shown promising electrochemical performance in the past flow battery study when used with different electrolyte conditions.11,13,16,29,30 We chose the five hydrotropes because they were structurally diverse, they did not contain redox-active functional groups, and they were previously shown to increase the solubilities of hydrophobic organic compounds in past studies.15,22,28,31

Figure 1.

Figure 1

Open in a new tab

(A) Five redox-active organic compounds and (B) five hydrotropes used in this study.

We examined that the organic molecules’ solubility changes in the presence of different concentrations of hydrotropes, ranging from 0 to 20 weight percent (wt %). Note that hydrotrope concentrations are conventionally expressed in units of weight percent in the literature.22 For the hydrotrope–compound combinations that yielded the highest solubility increases, we used experimental electrochemical tests to determine how hydrotropes influenced the compounds’ formal reduction potential (E0′), diffusivity of the electrochemically active species (D), and the standard rate constant (k0). We used a cyclic voltammetry fitting algorithm to model our experimental data and determined values for E0′, D, k0, and the transfer coefficient (α). We also conducted the linear sweep voltammetry and electrochemical impedance spectroscopy using a rotational disk electrode setup to experimentally confirm the diffusion coefficient and standard rate constant values determined from cyclic voltammetry fitting algorithm.

Results and Discussion

Influence of Hydrotropes on Aqueous Solubilities

The aqueous solubilities of the five redox-active compounds varied as a function of hydrotrope type and concentration, with larger solubility changes observed at higher hydrotrope concentrations (Figure 2). The aqueous solubilities of the redox-active compounds in the absence of a hydrotrope were similar to previously measured values, except for FMN, which was substantially more soluble in our study (1.3 M) than was previously reported (0.1 M) (Table S1).11,16,29,32 Note the background electrolytes used in our experiments differed depending on the redox-active compound, as published redox flow battery studies employing the compounds were operated at different pH values (4-OH-TEMPO and MeV: 1 M KCl, pH ≈ 7; AQDS and BDS: 1 M H2SO4, pH ≈ 0; FMN: 1 M KOH, pH ≈ 14).11,14,16,29 The solubility changes induced by a hydrotrope differed for each redox-active compound. For example, sodium xylenesulfonate (NaXS) significantly increased the solubility of 4-OH-TEMPO from 1.18 ± 0.04 M (0 wt %) to 1.99 ± 0.12 M (20 wt %), but NaXS significantly lowered the solubilities of AQDS from 0.83 ± 0.03 M (0 wt %) to 0.39 ± 0.06 M (20 wt %) and BDS from 0.52 ± 0.01 M (0 wt %) to 0.03 ± 0.05 M (20 wt %). In contrast, proline, another hydrotrope, exhibited the opposite trend; it increased the solubilities of BDS from 0.52 ± 0.01 M (0 wt %) to 0.88 ± 0.02 M (20 wt %) and of AQDS from 0.83 ± 0.03 M (0 wt %) to 1.05 ± 0.23 M (15 wt %), and it decreased the solubility of 4-OH-TEMPO (Figure 2A,D,E). Unlike the other redox-active compounds, MeV did not undergo substantial solubility changes in the presence of the hydrotropes (Figure 2B), which may be due to it having a higher aqueous solubility (2.69 ± 0.01 M) than the other compounds. Note we express the concentration of the hydrotrope using wt % due to the use of this convention in the hydrotrope literature.22 A figure containing hydrotrope concentrations expressed in terms of molarity is in the SI (Figure S1).

Figure 2.

Figure 2

Open in a new tab

Aqueous solubilities of redox-active compounds as a function of hydrotrope concentration and type. The following acronyms are used in the figure; MeV: methyl viologen, FMN: riboflavin-5′-monophosphate, AQDS: anthraquinone-2,7-disulfonate, BDS: 4,5-dihydroxy-1,3-benzenedisulfonate, NaXS: sodium xylenesulfonate, NapTS: sodium para-toluenesulfonate, and TBA: tert-butanol. Error bars represent the standard deviation for measurements made from triplicate reactors. Supporting electrolyte compositions: 1 M KCl (4-OH-TEMPO and MeV), 1 M KOH (FMN), and 1 M H2SO4 (AQDS and BDS).

The magnitude and direction of hydrotrope-induced solubility changes varied as a function of both the redox-active compound and the hydrotrope, which is consistent with observations made for sparingly soluble organic compounds solubilized by hydrotropes.27,33−37 The lack of a universal trend is likely a consequence of multiple factors. First, hydrotropes are known to interact with and solubilize organic compounds via specific molecular interactions, such as π–π interactions, with the strength of the specific attractive (or repulsive) forces depending on the chemical structures of the hydrotrope and compound.24,33 Second, the electrolyte composition and solution pH values likely influenced hydrotrope–compound interactions. We used different pH values for different redox-active compounds, which could affect the hydrogen bonding and alter the structures of hydrotropes that undergo acid–base reactions [NaXS pKA (for 3,4-dimethylbenzenesulfonate): 10.36;38 proline pKA,1: 2.0, pKA,2: 10.6;39]. Third, the solubilities of AQDS and BDS may have been decreased in the presence of NaXS or NapTS because both of these organic compounds are known to precipitate as sodium salts at low pH values.29

We observed that the largest solubility increases for (a) 4-OH-TEMPO in the presence of the hydrotropes NaXS, NapTS, and urea and (b) BDS in the presence of the hydrotrope proline (Figure 2). Consequently, we focused on these compound–hydrotrope combinations in our subsequent experiments to understand how hydrotropes that increase the solubilities of redox-active organic compounds influence the compounds’ formal reduction potential values, diffusion coefficients, and interfacial electron-transfer kinetics.

Influence of Hydrotropes on 4-OH-TEMPO Electrochemical Parameters

We used cyclic voltammetry to examine how the presence of NaXS, NapTS, and urea influenced E0′, D, and k0 of 4-OH-TEMPO (Figure 3A), which undergoes a single reversible electron-transfer reaction at circumneutral pH values to form an oxidized radical species (E0′ = 0.59 V vs Ag/AgCl,11Figure 3B). The presence of the hydrotrope did not influence E0′ values (no hydrotrope: 0.59 ± 0.00 V vs Ag/AgCl, 20 wt % NaXS: 0.59 ± 0.01 V, 20 wt % NapTS: 0.59 ± 0.00 V, 20 wt % urea: 0.59 ± 0.00 V), indicating that hydrotropes did not preferentially alter solvation energies (ΔGsolv0) of the reduced or oxidized forms of 4-OH-TEMPO.40 In control experiments lacking 4-OH-TEMPO, we confirmed that hydrotropes did not undergo any electron-transfer reactions over the same potential window (Figure S2). All four CVs were reversible at a scan rate of 10 mV/s. We calculated peak separation values, and they were close to the theoretical value for a reversible one-electron redox reaction at neutral pH (peak separation: no hydrotrope: 60 mV, 20 wt % NaXS: 62 mV, 20 wt % NapTS: 61 mV, 20 wt % urea: 63 mV), indicating that the standard reaction rate constant (k0) was sufficiently large relative to the diffusivities of both reduced and oxidized form of 4-OH-TEMPO (Dred and Dox) to achieve reversible electron-transfer kinetics in the presence and absence of hydrotropes.41 We found that peak currents were lower in the presence of a hydrotrope relative to the control condition lacking a hydrotrope (Figure 3A), which indicated that hydrotropes decreased the diffusivity of both 4-OH-TEMPO and the oxidized 4-OH-TEMPO species.

Figure 3.

Figure 3

Open in a new tab

(A) Cyclic voltammograms of 1 mM 4-OH-TEMPO in 1 M KCl, 20 wt % urea, 20 wt % NapTS, and 20 wt % NaXS at v = 10 mV/s. The aqueous portions of hydrotrope solutions contained 1 M KCl. (B) Half-reaction of 4-OH-TEMPO.

We collected CVs for each solution at several scan rates (0.001, 0.01, 0.05, 0.1, 0.5, and 1 V/s) to quantify the D and k0 values of 4-OH-TEMPO in the presence and absence of a hydrotrope. We background subtracted the capacitive current from voltammograms using data collected for otherwise identical solutions lacking the redox-active compound and fitted the CVs of all scan rates simultaneously to a finite-element model described in the experimental section (Figure 4). We confirmed the accuracies of the values determined from the CVs for a subset of the conditions by performing linear sweep voltammetry measurements (LSV) using a rotating disk electrode cell and electrochemical impedance spectroscopy (EIS). The main advantage of performing cyclic voltammetry experiments over rotating disk electrode experiments was that CV measurements could be made with far smaller sample volumes.

Figure 4.

Figure 4

Open in a new tab

CV experimental results (solid lines) and simulation fits (dashed lines) for 1 mM 4-OH-TEMPO in (A) 1 M KCl, (B) 20 wt % urea in 1 M KCl, (C) 20 wt % NapTS in 1 M KCl, and (D) 20 wt % NaXS in 1 M KCl. The best-fit simulation parameters are provided in Table S2.

We confirmed that the diffusion coefficient for 4-OH-TEMPO (Dred) was moderately larger in the absence of a hydrotrope (7.00 ± 0.24 × 10–6 cm2/s) than in cases in which 20 wt % hydrotrope was present (NaXS: 3.33 ± 0.04 × 10–6 cm2/s, NapTS: 4.27 ± 0.14 × 10–6 cm2/s, urea: 5.90 ± 0.07 × 10–6 cm2/s) from CV fitting results (Table 1). We observed the same trend of Dred values decreasing in the presence of a hydrotrope in the LSV data, although the absolute Dred values were all calculated to be slightly higher (no hydrotrope: 10.53 × 10–6 cm2/s, NaXS: 6.63 × 10–6 cm2/s, NapTS: 8.53 × 10–6 cm2/s, Table 1). The Dred values measured in the absence of a hydrotrope (CV: 7.00 ± 0.24 × 10–6 cm2/s, LSV: 10.53 × 10–6 cm2/s) both agree well with previously reported values in aqueous solutions (7.7 × 10–6,42 11.5 × 10–6,43 29.5 × 10–6 cm2/s11). The diffusion coefficients for the oxidized 4-OH-TEMPO radical (Dox) were calculated to be slightly smaller or approximately the same as the Dred values for each of the solutions tested (Table 1). There were not any significant differences between the E0′ or the transfer coefficient (α) values fitted from the CVs in the absence or presence of a hydrotrope (Table S2).

Table 1. Calculated Dred, Dox, and k0 Values for 4-OH-TEMPO Collected from CV, LSV, and EIS Experimentsa.

  Dred [cm2/s]
 Dox [cm2/s]
 k0 [cm/s]
electrolyte CV LSV CV LSV CV LSV EIS
1 M KCl 7.00 ± 0.24 × 10–6 10.53 × 10–6 6.82 ± 0.32 × 10–6 6.03 × 10–6 0.35 ± 0.14 0.09 ± 0.07 0.08 ± 0.00
20 wt % NaXS 3.33 ± 0.04 × 10–6 6.63 × 10–6 3.50 ± 0.11 × 10–6 3.76 × 10–6 0.48 ± 0.19 0.12 ± 0.11 0.29 ± 0.11
20 wt % NapTS 4.27 ± 0.14 × 10–6 8.53 × 10–6 4.92 ± 0.15 × 10–6 4.40 × 10–6 0.95 ± 0.07 0.32 ± 0.33 0.12 ± 0.01
20 wt % urea 5.90 ± 0.07 × 10–6   5.55 ± 0.12 × 10–6   0.78 ± 0.16    
Open in a new tab
a

The CV experiments were conducted with an initial 4-OH-TEMPO concentration of 1 mM. The LSV and EIS experiments were conducted with 1 mM 4-OH-TEMPO and 1 mM-oxidized 4-OH-TEMPO radical. The solutions contained 1 M KCl, 20 wt % NaXS in 1 M KCl, 20 wt % NapTS in 1 M KCl, or 20 wt % urea in 1 M KCl.

The observation that the presence of a hydrotrope decreased Dred values was likely a consequence of hydrotropes increasing the solution viscosity, as opposed to the hydrotrope actively altering the mobility of 4-OH-TEMPO. According to the Stokes–Einstein equation, which describes the movement of spherical particles, diffusion coefficients are inversely proportional to the solution viscosity

graphic file with name ao1c05133_m001.jpg 1

where kb is the Boltzmann constant, T is the thermodynamic temperature, μ is the dynamic viscosity of the solution, and r is the molecular radius. We measured the viscosities of each solution in the absence of a hydrotrope (9.1 × 10–4 Pa·s) and in the presence of 20 wt % NaXS (1.8 × 10–3 Pa·s) or NapTS: (1.9 × 10–3 Pa·s). Using these values, we calculated the hydraulic radius of 4-OH-TEMPO based on Dred from the CV data, μ, and T (298K) for the three solutions. The calculated hydraulic radii of 4-OH-TEMPO were similar for all of the solutions (no hydrotrope: 3.3 Å, NaXS: 3.7 Å, NapTS: 2.7 Å), indicating that the hydrotrope did not actively alter the hydrated radius of 4-OH-TEMPO via molecular interactions.

We quantified k0 values for 4-OH-TEMPO in each of the solutions by fitting the CV data collected with a glassy carbon electrode (Table 1). The data generally showed that the presence of a hydrotrope slightly increased the k0 values, although the differences were near the experimental errors. For a subset of the experimental conditions, we also quantified k0 values using the LSV data (Figures S3 and S4)44 and EIS measurements (Figure S5). The k0 values agreed well among the techniques, with the CV values being slightly larger than those obtained with the other two techniques (Table 1). The k0 value measured in the absence of a hydrotrope (CV value: 0.35 ± 0.14 cm/s) was approximately an order of magnitude higher than a previously reported value measured in an aqueous electrolyte using a glassy carbon working electrode (0.0157 cm/s).43 The reason for the discrepancy among the values measured here and the value reported in the previous study remains unclear.

No previous studies have examined how hydrotropes influence k0 values, but insights can be gained from comparing the observations reported here and the previous studies examining the effects that surfactants have on k0 values. In contrast to our observations, surfactants can have a large impact on k0 values, with changes scaling up to 2 orders of magnitude in either direction.45 In the case of the surfactants, the direction and scale of the effect appear to depend both on the nature and strength of electrostatic and hydrophobic interactions between the surfactant and solute.45−48 The lack of a substantial change in k0 values observed here suggests that 4-OH-TEMPO does not strongly interact with the hydrotropes studied, consistent with the hydrotropes having a minimal impact D or E0′ values.

Influence of Hydrotrope on BDS

We characterized the electrochemical properties of BDS (Figure 5A) in the presence of the hydrotrope proline to determine if the trends observed for 4-OH-TEMPO also applied to another hydrotrope redox-active compound combination. BDS coupled with hydrotrope proline yielded the largest solubility increase in BDS (Figure 2E). The compound BDS is known to undergo a two-electron-transfer reaction in acidic solutions (Figure 5B).29,49 The cyclic voltammograms collected in the presence and absence of proline demonstrated that proline did not alter E0′ (0.71 ± 0.01 V vs Ag/AgCl for both solutions, Figure 5B), indicating that proline did not preferentially alter the solvation energies (ΔGsolv0) of the reduced or oxidized forms of BDS.40 In a control CV experiment lacking the redox-active compound BDS, we confirmed that proline itself did not undergo any electron-transfer reaction (Figure S2). Note the E0′ value measured here in 1 M H2SO4 (0.70 V vs Ag/AgCl) was slightly higher than a previously reported E0′ value (0.64 V vs Ag/AgCl) measured in 1 M H2SO4.29,49 The CVs collected at a scan rate of 10 mV/s showed a quasi-reversible current response with large peak separations (peak separation: no hydrotrope: 251 mV, proline: 257 mV), indicating that k0 was not sufficiently large relative to Dox and Dred values to achieve fully reversible electron-transfer kinetics.41 The high-peak separation value aligned with a literature reported value at the same scan rate (280 mV).49 The decreases in the peak current heights in the presence of a hydrotrope also suggested that the hydrotrope decreased both Dox and Dred values.

Figure 5.

Figure 5

Open in a new tab

(A) Cyclic voltammetry of 1 mM 4,5-dihydroxy-1,3-benzenedisulfonate (BDS) in 1 M H2SO4 and 1 M H2SO4 with 15 wt % proline. v = 10 mV/s. (B) Half-reaction of BDS.

We used the same approach as for 4-OH-TEMPO, in which we simultaneously fit cyclic voltammograms collected at several scan rates to determine D values for BDS in the presence and absence of proline (Figure 6). The diffusion coefficient for BDS (Dred) in the absence of a hydrotrope (3.85 ± 0.21 × 10–6 cm2/s) was larger than the value in 15 wt % proline (1.97 ± 0.04 × 10–6 cm2/s) (Table 2). The LSV data corroborated this trend, although Dred values were approximately twice as large when calculated from LSV data (no hydrotrope: Dred,LSV: 4.88 × 10–6 cm2/s, proline: Dred,LSV: 2.33 × 10–6 cm2/s) (Table 2). In addition, we observed in the CV simulation results that in both solutions, Dox was smaller than Dred. The difference between BDS’s Dred and Dox was reflected on the voltammograms’ peak currents in each scan direction, where the peak current in the oxidation side was higher than the value of the peak current in the reduction side (Figure 5A). When fitting CV curves, we tested if a chemical degradation reaction of oxidized BDS caused the lower peak current during reduction relative to the oxidation peak, but adding a chemical reaction to the model did not improve the quality of the fit or the fitted electrochemical parameters (Figure S6, Table S3). The Dred values we measured in 1 M H2SO4 (Dred, CV: 3.85 ± 0.21 × 10–6 cm2/s, Dred, LSV: 4.88 × 10–6 cm2/s) were both comparable to previously reported values in 1 M H2SO4 (5.1 × 10–6,49 3.8 × 10–6 cm2/s13). We did not observe a significant difference between the E0′ and α value from the fitting results with and without the presence of a hydrotrope (Table S3). We confirmed that the reason D decreased in the presence of the hydrotrope was most likely a consequence of proline increasing the solution viscosity, as the radii of BDS in both solutions calculated with eq 1 were approximately equal (no hydrotrope: 5.8 Å, proline: 6.6 Å).

Figure 6.

Figure 6

Open in a new tab

CV simulation and experiment result of 1 mM BDS in (A) 1 M H2SO4 and (B) 15 wt % proline in 1 M H2SO4.

Table 2. CV Simulation Fitting Results of 1 mM BDS in 1 M H2SO4 and 15 wt % Proline Solutionsa.

  Dred [cm2/s]
 Dox [cm2/s]
 k0 [cm/s]
electrolyte CV LSV CV LSV CV LSV
1 M H2SO4 3.85 ± 0.21 × 10–6 4.88 × 10–6 3.67 ± 0.86 × 10–6   1.73 ± 0.10 × 10–5 20.42 ± 8.53 × 10–5
15 wt % proline 1.97 ± 0.04 × 10–6 2.33 × 10–6 2.94 ± 0.15 × 10–6   2.06 ± 0.07 × 10–5 9.84 ± 3.63 × 10–5
Open in a new tab
a

LSV fitting results of 2 mM BDS in 1 M H2SO4 and 15 wt % proline solutions in 1 M H2SO4.

We quantified k0 values for BDS using the CV and LSV data (Table 2, Figures S7 and S8). The CV data yielded the similar k0 values in the presence and the absence of proline (no hydrotrope: 1.73 ± 0.10 × 10–5 cm/s, 15 wt % proline: 2.06 ± 0.07 × 10–5 cm/s; Table 2), while from our LSV experiment, BDS’s k0 was found to decrease from 20.4 ± 8.5 × 10–5 cm/s in the absence of a hydrotrope to 9.8 ± 3.6 × 10–5 cm/s in 15 wt % proline. The reaction rate constant value reported for quinone compounds in 1 M H2SO4 was in the range of 10–3–10–4 cm/s,13,29 which was close to our calculated heterogeneous rate constant (kh) value of BDS in the absence of a hydrotrope from the LSV data (1.52 ± 0.46 × 10–4 cm/s). Collectively, the data indicated that proline slightly decreased k0 on a glassy carbon electrode.

Implications for Flow Batteries

The solubility experiments indicated that hydrotropes can increase the solubilities of redox-active compounds relevant to aqueous redox flow batteries. There was not a general trend in how each hydrotrope influenced the solubilities of the redox-active compounds, as each compound–hydrotrope combination yielded different degree of solubility change, consistent with past studies examining the solubilization of organic compounds using hydrotropes.27,33−37 We successfully demonstrated that the hydrotrope NaXS could increase the solubility of 4-OH-TEMPO by 69%, indicating that the theoretical energy storage density of the aqueous redox flow battery using 4-OH-TEMPO as the electroactive species could be significantly improved. Our electrochemical data indicated that for 4-OH-TEMPO and BDS, the hydrotropes tested had only a minor impact on their electrochemical properties. For both 4-OH-TEMPO and BDS, the addition of a hydrotrope slightly decreased both Dox and Dred values, which was likely caused by the hydrotropes increasing solution viscosities. With respect to k0, 4-OH-TEMPO’s value slightly increased in the presence of hydrotropes, while the BDS’s value slightly decreased. Compared to molecular modification techniques, amending the solution with hydrotropes had less of an effect on a redox-active compound solubility,50 but it offered an alternative means for achieving solubility changes without substantially altering the reactivities of the compound. On a practical level, the presence of a hydrotrope will increase the solution viscosity, which would increase pumping costs in a flow battery.

Experimental Section

Chemicals

All of the chemicals used were used without further purification. The five redox-active organic molecules studied were 4-OH-TEMPO free radical (>98%, Alfa Aesar), methyl viologen (98%, Acros Organics), riboflavin-5′-monophosphate (>93%, Tokyo Chemical Industry Co.), anthraquinone-2,7-disulfonate (>97%, Tokyo Chemical Industry Co.), and 4,5-dihydroxy-1,3-benzenedisulfonate (97%, Sigma-Aldrich). The five hydrotropes used were proline (99%, Alfa Aesar), sodium xylenesulfonate (NaXS) (<9.0% sodium sulfate, Sigma-Aldrich), sodium para-toluenesulfonate (NapTS) (99%, Spectrum), urea (GR for analysis, Sigma-Aldrich), and tert-butanol (TBA) (≥99.0%, Sigma-Aldrich). We electrochemically oxidized 4-OH-TEMPO radical (E0′ = 0.59 V vs Ag/AgCl) to 4-OH-TEMPO+ for use in electrochemical tests with a custom-made glassy carbon bulk electrolysis cell by applying a potential of 0.75 V vs Ag/AgCl (MF-2056, BASi) for 5–10 h, depending on the solution concentration. The conversion of 4-OH-TEMPO radical to 4-OH-TEMPO+ was monitored by stopping the electrochemical oxidation reaction once per hour and measuring the open-circuit potential. It was stopped when the open-circuit potential reached 0.72 V vs Ag/AgCl, which indicated that 99% of the 4-OH-TEMPO radical in the solution had been converted to 4-OH-TEMPO+ according to the Nernst equation

graphic file with name ao1c05133_m002.jpg 2

where E is the cell potential, E0 is the formal reduction potential, R is the universal gas constant, n is the number of electrons transferred, F is the Faraday constant, T is the temperature, and Q is the reaction quotient (in this case: Inline graphic).

Solubility Measurements

All experiments were conducted with deionized water having a resistivity >18 MΩ·cm. The solubility experiments were all performed in triplicate, with reported errors representing the standard deviations among them. The hydrotropic solvents were made gravimetrically by combining the hydrotrope salt or liquid with 1 M KCl, 1 M H2SO4, or 1 M KOH, depending on the organic compounds used. The maximum hydrotrope weight percentage was limited to 20 wt % because in the preliminary experiments, we observed phase separations when using higher hydrotrope concentrations and 1 M KCl. All hydrotrope solutions were equilibrated at least 24 h prior to the solubility experiments. After equilibration, an excess of a redox-active compound was added, allowed the solutions to equilibrate for at least 24 h, and then measured the concentration. The solubility tests on AQDS were not performed with the hydrotrope urea or TBA because we ran out of 97% purity AQDS compounds, and it was no longer commercially available.

The solubilities of each organic compound in each solution were measured using UV–vis spectroscopy (PerkimElmer Lambda35 UV–vis Spectrophotometer; UV-1800, Shimadzu USA MFG INC.). The PerkimElmer Lambda35 UV–vis was used for experiments with NaXS, NapTS, and proline, and the UV-1800 instrument was used for experiments with urea and TBA. The reason for the switch between instruments was that the Lambda35 stopped working over the course of the experiments. For all solubility measurements, a small volume (∼10 μL) from the top layer of each saturated solution without any visible undissolved suspended solids presented was taken, then got diluted 10 000-fold with DI water prior to measurements to decrease the absorbance to a value within our standard curves. For a subset of the solutions, the initial sample was filtered through a 0.2 μm PTFE syringe filter prior to the dilution step to rule out the possibility that the samples contained suspended particulate material. Filtering the samples had a negligible effect on the measured solubility values. The dilution step decreased the hydrotrope concentration in our measured solutions, which was important because some hydrotropes had overlapping absorbances with some of the redox-active compounds. No hydrotrope peaks were present in the UV–vis spectra in control measurements lacking the redox-active compounds. The wavelengths used to record peak absorbances for each compound were as follows: MeV: 258 nm, 4-OH-TEMPO: 243 nm, AQDS: 328 nm, BDS: 291 nm, and FMN: 446 nm. The molar attenuation coefficient for each compound was calculated from our standard curves using the Beer–Lambert Law (MeV: 1.35 ± 0.17 × 103 m2/mol, 4-OH-TEMPO: 2.45 ± 0.10 × 102 m2/mol, AQDS: 6.04 ± 0.21 × 102 m2/mol, BDS: 3.50 ± 0.33 × 102 m2/mol, and FMN: 7.30 ± 0.85 × 102 m2/mol).51 The calibration curves for each compound are provided in the Supporting Information (Figure S9).

Cyclic Voltammetry

Cyclic voltammetry experiments were conducted using a CH Instruments potentiostat (model 630E) in a three-electrode cell containing a glassy carbon working electrode (BASi MF-2012, 3 mm diameter, area = 0.0707 cm2), an Ag/AgCl reference electrode in 3 M KCl (BASi MF-2056), and a platinum wire counter electrode. Before each experiment, the working electrode was polished with 0.05 μm of alumina particles on the microcloth pad and then rinsed with DI water and methanol. For 4-OH-TEMPO, the pH of all electrolytes containing a hydrotrope was readjusted to pH = 7.0 using 1 M HCl solution. For BDS, the pH of the electrolytes containing a hydrotrope was readjusted to ∼0 using 2.5 M H2SO4. For each organic compound, the triplicate CVs were collected in scan rates of 0.001, 0.01, 0.05, 0.1, 0.5, and 1 V/s. For each scan rate, three cycles of CV data were collected. The CV data without the redox-active organic compound in otherwise identical electrolytes using the same scan rates were also collected to subtract out the background capacitive current from data.

A finite-element model describing the reactions during cyclic voltammetry was created using the method described in Bard and Faulkner.41 1D finite-element analysis was used with uniform space (Δx) and time (Δt) discretization to solve the concentrations of the oxidized (co) and reduced (cr) species as functions of time and distance from the electrode. The concentrations followed Fick’s second law of diffusion in one dimension

graphic file with name ao1c05133_m004.jpg 3

where Dm is the diffusion coefficient of species m. For the boundary condition farthest from the electrode, xn,max, we held the concentration constant at the starting bulk concentration; we modeled a sufficiently large distance from the electrode (>0.5 cm) such that this boundary condition had no impact on the simulated voltammograms. For the electrode boundary condition, x0, the electrochemical reactions at the electrode surface were modeled as first-order reactions. The rate constants for the forward (reduction) reaction and backward (oxidation) reaction, kf and kb, respectively, were defined as in the Butler–Volmer equation

graphic file with name ao1c05133_m005.jpg 4
graphic file with name ao1c05133_m006.jpg 5

where k0 is the electrochemical rate constant (cm/s), α is the unitless charge-transfer coefficient, n is the number of electrons transferred, E(t) is the electrode potential at time t, E0′ is the formal reduction potential, and F is Faraday’s constant. We solved Fick’s first law of diffusion to obtain the flux of species m at the surface, Jm(t)

graphic file with name ao1c05133_m007.jpg 6

Given the flux of the oxidized species at the electrode, the current due to the electrochemical reaction Inline graphic(t) was defined by Faraday’s law of electrolysis

graphic file with name ao1c05133_m009.jpg 7

The CV simulation was written as a function with the following inputs: the diffusion coefficients of the oxidized and reduced species (Do, Dr), E0′, k0, α, uniformly spaced E(t), scan rate (ν), and total concentration. It was assumed that for a voltammogram that started with an increasing electrode potential, the starting concentration of the reduced species for all space was equal to the total concentration. The simulation then solved the set of differential equations to return the current density (mA/cm2) for three consecutive cycles. The third cyclic voltammogram was compared to the experimentally collected third scan of the cyclic voltammogram after background subtracting the capacitive current that was experimentally measured under identical conditions.

The simulation was regressed through repeated cycles to find the values of Do, Dr, E0′, k0, and α whose simulation current, î(t), best fit the background-subtracted experimental current, i(t), data. The data were fitted for all scan rates simultaneously using a single objective function. The best-fit values for Do, Dr, E0′, k0, and α were those that minimized the weighted sum of a squares error function, WSSE

graphic file with name ao1c05133_m010.jpg 8

where the weight, w, at time tj was defined as the following

graphic file with name ao1c05133_m011.jpg 9

In eq 9, Inline graphic refers to the maximum value of variable X among the set of time points tk that have the same scan rate as the point tj. For example, if tj was obtained from a CV with a scan rate 0.05 V/s, Inline graphic would be the maximum absolute value of the current from that CV. Conceptually, this weighting function is the product of (a) the inverse of the absolute value of maximum current obtained for that scan rate and (b) the normalized second derivative of current with respect to voltage at that time point. Part a of the weighting factor accounts for the increase in the current magnitudes with the scan rate, preventing it from overprioritizing the quality of fit to higher scan rates, where the same percent error yields a larger sum of squares error. Part b of the weighting factor prioritizes the quality of fit to the peaks in the voltammogram, which are the most characteristic features in a CV with a Faradaic reaction. The second derivative at a given time index j is approximated by fitting the 15 points around it in the same scan direction {j – 7, j – 6, ..., j + 7} to a parabola and taking the second derivative of the parabola. The WSSE objective function was minimized using a genetic algorithm (GA package v.3.2 in R v.4.0.3).52,53 Convergence was defined as 20 generations where the minimum value is unchanged or after a maximum of 80 generations.

The uncertainties of fitting parameters were approximated using a jack-knifing method that treated each scan rate as an independent sample. Instead of regressing the simulation to the entire data set, we regressed it to the partial data set, where one scan rate was removed. This was repeated for all possible partial data sets in which only one scan rate was removed, producing a set of fitting parameters. The standard deviation of this set of fitting parameters was treated as an estimate for the standard error of the best fit.54

For BDS, the simulation was adjusted to include a term for the decomposition of the oxidized species to determine if including degradation improved the model’s quality of fit. The degradation was modeled as a first-order reaction

graphic file with name ao1c05133_m014.jpg 10

where co*(x,t) is the modified value of the concentration in the absence of a reaction, co(x,t), to include the spontaneous reaction. In these trials, including this term did not significantly improve the quality of the fits and did not significantly change the values for Do, Dr, E0′, k0, α, so we did not include it in the reported best-fit parameters.

Electrochemical Impedance Spectroscopy and Linear Sweep Voltammetry

Electrochemical impedance spectroscopy (EIS) and linear sweep voltammetry (LSV) experiments were conducted with Gamry potentiostat (REF 600P) in a three-electrode cell: a glassy carbon rotation disk electrode (5 mm diameter Teflon-encased glassy carbon disk, Pine Research Instrument), a Ag/AgCl reference electrode (MF-2056, BASi), and a platinum wire counter electrode. Before each experiment, the electrode was polished using the same protocol as described for cyclic voltammetry experiments. For EIS and LSV experiments with 4-OH-TEMPO, the solutions used were 1 mM radical 4-OH-TEMPO and 1 mM 4-OH-TEMPO+ dissolved in 1 M KCl, 20 wt % NaXS, or 20 wt % NapTS. The known concentrations of both 4-OH-TEMPO species facilitated the calculation of exchange current density

graphic file with name ao1c05133_m015.jpg 11

where Rct is the charge-transfer resistance in Ω and j0 is the exchange current density in A/cm2 (electrode area: 0.196 cm2). The EIS experiment for 4-OH-TEMPO was conducted in an open-circuit potential, with a rotation rate of 2500 rpm, an amplitude of 10 mV, and frequencies between 0.5 and 100 000 Hz. The Nyquist plot was fitted using Electrochemical Impedance Spectroscopy Software from Gamry to quantify Rct, which was then used to calculate j0 according to eq 11. Then, j0 was used to calculate the electrochemical rate constant (k0) assuming that α was 0.5 (note that the α values from our CV simulation results were close to 0.5; Tables S3 and S4)

graphic file with name ao1c05133_m016.jpg 12

The LSV experiments for 4-OH-TEMPO were carried out at a scan rate of 10 mV/s, and the electrode was rotated at 1500, 1750, 2000, 2250, and 2500 rotations per minute (rpm). The cutoff voltages were set to 0.3 and 0.9 V vs Ag/AgCl. The oxidizing and reducing limiting currents were measured at 0.88 and 0.30 V, respectively. The kinetically dominated currents were selected at overpotentials than 0.05 V. Both the limiting current and the kinetically dominated current were fitted against reciprocal of the square root of rotation rate (rad/s) according to the Koutecky–Levich equation44

graphic file with name ao1c05133_m017.jpg 13

where D is the diffusion coefficient, ν is the kinematic viscosity, C is the bulk concentration, ω is the rotation rate, η is the overpotential, and kh is the heterogeneous rate constant for electron transfer, defined as

graphic file with name ao1c05133_m018.jpg 14

The first half of eq 13 describes the mass transfer dominated limiting currents, and the second half describes the kinetically dominated (cathodic and anodic) currents. The current was fitted using eq 13 to yield a linear plot, with the slope of Inline graphic and intercept of Inline graphic, where r was 0.25 mm. By fitting limiting currents and kinetically dominated currents against the reciprocal of the square root of the rotation rate, D and kh values of both radical 4-OH-TEMPO and 4-OH-TEMPO+ were calculated from the slope and intercept values.

For BDS, LSV experiments were conducted using solutions containing 2 mM BDS in 1 M H2SO4 and 15 wt % proline. The rotation rates applied for BDS’s LSV experiments were 500, 750, 1000, 1250, and 1500 rpm. Because the solution started with only the reduced species, we only recorded oxidizing limiting currents at 1.15 V and kinetically dominated anodic currents at overpotential <0.05 V. The diffusion coefficient of reduced BDS species was calculated using eq 13 using the same approach as described above for 4-OH-TEMPO.

Viscosity Measurements

The viscosities of 1 M KCl, 20 wt % NaXS, and 20 wt % NapTS were measured using a rotational rheometer (RFS3) with concentric cylinder geometry with a shear rate from 10 to 100/s (Table S4). The viscosities of 1 M H2SO4 and 15 wt % proline were measured using a Ubbelohde glass viscometer (CUC-50 CANNON-Ubbelohde calibrated) (Table S4).

Acknowledgments

The authors acknowledge the financial support for this research from Penn State University SEED Grant program and National Science Foundation Grant no. CBET-1603635. The authors thank Aijie Han’s assistance of the viscosity measurements for NaXS, NapTS, and proline’s solution.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.1c05133.

  • The comparison of solubilities of redox-active compounds from our experimental measurements and reference values in the literature; the aqueous solubilities of redox-active compounds as a function of hydrotrope concentration in molarity; cyclic voltammograms for background electrolytes without the redox-active organic compound; the fitting results of 4-OH-TEMPO CVs; linear sweep voltammograms of 1 mM 4-OH-TEMPO(+) and 1 mM 4-OH-TEMPO(•); Koutecký–Levich plots for 1 mM 4-OH-TEMPO(+) and 1 mM 4-OH-TEMPO(•); Nyquist plots from EIS measurements; comparison of CV simulation results with and without the chemical degradation parameter; the fitting results of BDS CVs; linear sweep voltammograms of 2 mM 4,5-dihydroxy-1,3-benzenedisulfonate (BDS); Koutecký–Levich plots for 2 mM BDS; calibration curves (UV–vis absorbance versus concentration) for the five redox-active organic compounds; and the hydrodynamic viscosity measurements of 1 M KCl, 20 wt % NaXS, and 20 wt % NapTS, 1 M H2SO4, and 15 wt % proline (PDF)

The authors declare no competing financial interest.

Supplementary Material

ao1c05133_si_001.pdf (1.4MB, pdf)

References

  1. Wang W.; Luo Q.; Li B.; Wei X.; Li L.; Yang Z. Recent Progress in Redox Flow Battery Research and Development. Adv. Funct. Mater. 2013, 23, 970–986. 10.1002/adfm.201200694. [DOI] [Google Scholar]
  2. Crawford A.; Viswanathan V.; Stephenson D.; Wang W.; Thomsen E.; Reed D.; Li B.; Balducci P.; Kintner-Meyer M.; Sprenkle V. Comparative Analysis for Various Redox Flow Batteries Chemistries Using a Cost Performance Model. J. Power Sources 2015, 293, 388–399. 10.1016/j.jpowsour.2015.05.066. [DOI] [Google Scholar]
  3. Nourai A. In Large-Scale Electricity Storage Technologies for Energy Management, Proceedings of IEEE Power Engineering Society Transmission and Distribution Conference, 2002; pp 310–315.
  4. Soloveichik G. L. Flow Batteries: Current Status and Trends. Chem. Rev. 2015, 115, 11533–11558. 10.1021/cr500720t. [DOI] [PubMed] [Google Scholar]
  5. Perry M. L.; Weber A. Z. Advanced Redox-Flow Batteries: A Perspective. J. Electrochem. Soc. 2016, 163, A5064–A5067. 10.1149/2.0101601jes. [DOI] [Google Scholar]
  6. Bachman J. E.; Curtiss L. A.; Assary R. S. Investigation of the Redox Chemistry of Anthraquinone Derivatives Using Density Functional Theory. J. Phys. Chem. A 2014, 118, 8852–8860. 10.1021/jp5060777. [DOI] [PubMed] [Google Scholar]
  7. Leung P.; Shah A. A.; Sanz L.; Flox C.; Morante J. R.; Xu Q.; Mohamed M. R.; Ponce de León C.; Walsh F. C. Recent Developments in Organic Redox Flow Batteries: A Critical Review. J. Power Sources 2017, 360, 243–283. 10.1016/j.jpowsour.2017.05.057. [DOI] [Google Scholar]
  8. Li B.; Liu J. Progress and Directions in Low-Cost Redox-Flow Batteries for Large-Scale Energy Storage. Natl. Sci. Rev. 2017, 4, 91–105. 10.1093/nsr/nww098. [DOI] [Google Scholar]
  9. Li Z.; Lu Y. C. Redox Flow Batteries: Want More Electrons? Go Organic!. Chem 2018, 4, 2020–2021. 10.1016/j.chempr.2018.08.032. [DOI] [Google Scholar]
  10. Yang B.; Hoober-Burkhardt L.; Wang F.; Surya Prakash G. K.; Narayanan S. R. An Inexpensive Aqueous Flow Battery for Large-Scale Electrical Energy Storage Based on Water-Soluble Organic Redox Couples. J. Electrochem. Soc. 2014, 161, A1371–A1380. 10.1149/2.1001409jes. [DOI] [Google Scholar]
  11. Liu T.; Wei X.; Nie Z.; Sprenkle V.; Wang W. A Total Organic Aqueous Redox Flow Battery Employing a Low Cost and Sustainable Methyl Viologen Anolyte and 4-HO-TEMPO Catholyte. Adv. Energy Mater. 2015, 6, 1501449 10.1002/aenm.201501449. [DOI] [Google Scholar]
  12. Huskinson B.; Marshak M. P.; Suh C.; Er S.; Gerhardt M. R.; Galvin C. J.; Chen X.; Aspuru-Guzik A.; Gordon R. G.; Aziz M. J. A Metal-Free Organic-Inorganic Aqueous Flow Battery. Nature 2014, 505, 195–198. 10.1038/nature12909. [DOI] [PubMed] [Google Scholar]
  13. Yang B.; Hoober-Burkhardt L.; Wang F.; Prakash G. K. S.; Narayanan S. R. An Inexpensive Aqueous Flow Battery for Large-Scale Electrical Energy Storage Based Onwater-Soluble Organic Redox Couples. J. Electrochem. Soc. 2014, 161, A1371. 10.1149/2.1001409jes. [DOI] [Google Scholar]
  14. Wedege K.; Dražević E.; Konya D.; Bentien A. Organic Redox Species in Aqueous Flow Batteries: Redox Potentials, Chemical Stability and Solubility. Sci. Rep. 2016, 6, 39101 10.1038/srep39101. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. Ding Y.; Zhang C.; Zhang L.; Wei H.; Li Y.; Yu G. Insights into Hydrotropic Solubilization for Hybrid Ion Redox Flow Batteries. ACS Energy Lett. 2018, 3, 2641–2648. 10.1021/acsenergylett.8b01828. [DOI] [Google Scholar]
  16. Orita A.; Verde M. G.; Sakai M.; Meng Y. S. A Biomimetic Redox Flow Battery Based on Flavin Mononucleotide. Nat. Commun. 2016, 7, 13230 10.1038/ncomms13230. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Ding Y.; Yu G. Molecular Engineering Enables Better Organic Flow Batteries. Chem 2017, 3, 917–919. 10.1016/j.chempr.2017.11.010. [DOI] [Google Scholar]
  18. DeBruler C.; Hu B.; Moss J.; Liu X.; Luo J.; Sun Y.; Liu T. L. Designer Two-Electron Storage Viologen Anolyte Materials for Neutral Aqueous Organic Redox Flow Batteries. Chem 2017, 3, 961–978. 10.1016/j.chempr.2017.11.001. [DOI] [Google Scholar]
  19. Schwan S.; Schröder D.; Wegner H. A.; Janek J.; Mollenhauer D. Substituent Pattern Effects on the Redox Potentials of Quinone-Based Active Materials for Aqueous Redox Flow Batteries. ChemSusChem 2020, 13, 5480–5488. 10.1002/cssc.202000454. [DOI] [PMC free article] [PubMed] [Google Scholar]
  20. Er S.; Suh C.; Marshak M. P.; Aspuru-Guzik A. Computational Design of Molecules for an All-Quinone Redox Flow Battery. Chem. Sci. 2015, 6, 885–893. 10.1039/c4sc03030c. [DOI] [PMC free article] [PubMed] [Google Scholar]
  21. Luo J.; Hu B.; Debruler C.; Liu T. L. A π-Conjugation Extended Viologen as a Two-Electron Storage Anolyte for Total Organic Aqueous Redox Flow Batteries. Angew. Chem. Int. Ed. 2018, 57, 231–235. 10.1002/anie.201710517. [DOI] [PubMed] [Google Scholar]
  22. Hodgdon T. K.; Kaler E. W. Hydrotropic Solutions. Curr. Opin. Colloid Interface Sci. 2007, 12, 121–128. 10.1016/j.cocis.2007.06.004. [DOI] [Google Scholar]
  23. Shimizu S.; Nagai Y. Effect of Solute Aggregation on Solubilization. J. Mol. Liq. 2019, 274, 209–214. 10.1016/j.molliq.2018.10.102. [DOI] [Google Scholar]
  24. Booth J. J.; Abbott S.; Shimizu S. Mechanism of Hydrophobic Drug Solubilization by Small Molecule Hydrotropes. J. Phys. Chem. B 2012, 116, 14915–14921. 10.1021/jp309819r. [DOI] [PubMed] [Google Scholar]
  25. Srinivas V.; Rodley G. A.; Ravikumar K.; Robinson W. T.; Turnbull M. M. Molecular Organization in Hydrotrope Assemblies. Langmuir 1997, 13, 3235–3239. 10.1021/la9609229. [DOI] [Google Scholar]
  26. Hopkins Hatzopoulos M.; Eastoe J.; Dowding P. J.; Rogers S. E.; Heenan R.; Dyer R. Are Hydrotropes Distinct from Surfactants?. Langmuir 2011, 27, 12346–12353. 10.1021/la2025846. [DOI] [PubMed] [Google Scholar]
  27. Coffman R. E.; Kildsig D. O. Hydrotropic Solubilization Mechanistic Studies. Pharm. Res. 1996, 13, 1460. 10.1023/a:1016011125302. [DOI] [PubMed] [Google Scholar]
  28. Coffman R. E.; Kildsig D. O. Effect of Nicotinamide and Urea on the Solubility of Riboflavin in Various Solvents. J. Pharm. Sci. 1996, 85, 951–954. 10.1021/js960012b. [DOI] [PubMed] [Google Scholar]
  29. Yang B.; Hoober-Burkhardt L.; Krishnamoorthy S.; Murali A.; Surya Prakash G. K.; Narayanan S. R. High-Performance Aqueous Organic Flow Battery with Quinone-Based Redox Couples at Both Electrodes. J. Electrochem. Soc. 2016, 163, A1442–A1449. 10.1149/2.1371607jes. [DOI] [Google Scholar]
  30. Pan F.; Wang Q. Redox Species of Redox Flow Batteries: A Review. Molecules 2015, 20, 20499–20517. 10.3390/molecules201119711. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Subramanian D.; Anisimov M. A. Phase Behavior and Mesoscale Solubilization in Aqueous Solutions of Hydrotropes. Fluid Phase Equilib. 2014, 362, 170–176. 10.1016/j.fluid.2013.09.064. [DOI] [Google Scholar]
  32. Permatasari A.; Lee W.; Kwon Y. Performance Improvement by Novel Activation Process Effect of Aqueous Organic Redox Flow Battery Using Tiron and Anthraquinone-2,7-Disulfonic Acid Redox Couple. Chem. Eng. J. 2020, 383, 123085 10.1016/j.cej.2019.123085. [DOI] [Google Scholar]
  33. Booth J. J.; Omar M.; Abbott S.; Shimizu S. Hydrotrope Accumulation around the Drug: The Driving Force for Solubilization and Minimum Hydrotrope Concentration for Nicotinamide and Urea. Phys. Chem. Chem. Phys. 2015, 17, 8028–8037. 10.1039/c4cp05414h. [DOI] [PubMed] [Google Scholar]
  34. Kim J. Y.; Kim S.; Papp M.; Park K.; Pinal R. Hydrotropic Solubilization of Poorly Water-Soluble Drugs. J. Pharm. Sci. 2010, 99, 3953–3965. 10.1002/jps.22241. [DOI] [PubMed] [Google Scholar]
  35. Da Silva R. C.; Spitzer M.; Da Silva L. H. M.; Loh W. Investigations on the Mechanism of Aqueous Solubility Increase Caused by Some Hydrotropes. Thermochim. Acta 1999, 328, 161–167. 10.1016/S0040-6031(98)00637-6. [DOI] [Google Scholar]
  36. Balasubramanian D.; Srinivas V.; Gaikar V. G.; Sharma M. M. Aggregation Behavior of Hydrotropic Compounds in Aqueous Solution. J. Phys. Chem. A. 1989, 93, 3865–3870. 10.1021/j100346a098. [DOI] [Google Scholar]
  37. Bauduin P.; Renoncourt A.; Kopf A.; Touraud D.; Kunz W. Unified Concept of Solubilization in Water by Hydrotropes and Cosolvents. Langmuir 2005, 21, 6769–6775. 10.1021/la050554l. [DOI] [PubMed] [Google Scholar]
  38. Duarte F.; Geng T.; Marloie G.; Al Hussain A. O.; Williams N. H.; Kamerlin S. C. L. The Alkaline Hydrolysis of Sulfonate Esters: Challenges in Interpreting Experimental and Theoretical Data. J. Org. Chem. 2014, 79, 2816–2828. 10.1021/jo402420t. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. O’Neil M. J.The Merck Index, 15th ed.; The Royal Society of Chemistry: Cambridge, 2013. [Google Scholar]
  40. Evans D. H. One-Electron and Two-Electron Transfers in Electrochemistry and Homogeneous Solution Reactions. Chem. Rev. 2008, 108, 2113–2144. 10.1021/cr068066l. [DOI] [PubMed] [Google Scholar]
  41. Bard A. J.; Faulkner L. R. Fundamentals and Applications. Electrochem. Methods 2001, 2, 580–632. [Google Scholar]
  42. Carlson E. D.; Majda M. Electrochemistry of TEMPO: An Assessment of the Water Diffusion Constant in the Aqueous Liquid/Vapor Interfacial Region. J. Solid State Electrochem. 2013, 3083–3091. 10.1007/s10008-013-2191-3. [DOI] [Google Scholar]
  43. Fan H.; Zhang J.; Ravivarma M.; Li H.; Hu B.; Lei J.; Feng Y.; Xiong S.; He C.; Gong J.; Gao T.; Song J. Radical Charge Population and Energy: Critical Role in Redox Potential and Cycling Life of Piperidine Nitroxyl Radical Cathodes in Aqueous Zinc Hybrid Flow Batteries. ACS Appl. Mater. Interfaces 2020, 12, 43568–43575. 10.1021/acsami.0c09941. [DOI] [PubMed] [Google Scholar]
  44. Treimer S.; Tanga A.; Johnson D. C. A Consideration of the Application of Koutecky-Levich Plots in the Diagnoses of Charge-Transfer Mechanisms at Rotated Disk Electrodes. Electroanalysis 2002, 14, 165–171. 10.1002/1521-4109(200202)14:33.0.co;2-6. [DOI] [Google Scholar]
  45. Samiey B.; Cheng C. H.; Wu J. Effects of Surfactants on the Rate of Chemical Reactions. J. Chem. 2014, 2014, 1–14. 10.1155/2014/908476. [DOI] [Google Scholar]
  46. Rajanna K. C.; Reddy K. N.; Kumar U. U.; Sai Prakash P. K. A Kinetic Study of Electron Transfer from L-Ascorbic Acid to Sodium Perborate and Potassium Peroxy Disulphate in Aqueous Acid and Micellar Media. Int. J. Chem. Kinet. 1996, 28, 153–164. . [DOI] [Google Scholar]
  47. Subba Rao P. V.; Krishna Rao G.; Ramakrishna K.; Rambabu G.; Satyanarayana A. Kinetics of Some Electron-Transfer Reactions of Iron(III)-2,2′-Bipyridyl Complex. Micellar Effect of Sodium Dodecyl Sulphate. Int. J. Chem. Kinet. 1997, 29, 171–179. . [DOI] [Google Scholar]
  48. Rodríguez A.; Del Mar Graciani M.; Balahura R.; Moyá M. L. Micellar Effects on the Electron Transfer Reaction within the Ion Pair [(NH3)5Co(N-Cyanopiperidine)]3+/[Fe(CN) 6]4-. J. Phys. Chem. B. 1996, 100, 16978–16983. 10.1021/jp961175r. [DOI] [Google Scholar]
  49. Rubio-Garcia J.; Kucernak A.; Parra-Puerto A.; Liu R.; Chakrabarti B. Hydrogen/Functionalized Benzoquinone for a High-Performance Regenerative Fuel Cell as a Potential Large-Scale Energy Storage Platform. J. Mater. Chem. A 2020, 8, 3933–3941. 10.1039/c9ta12396b. [DOI] [Google Scholar]
  50. VanGelder L. E.; Petel B. E.; Nachtigall O.; Martinez G.; Brennessel W. W.; Matson E. M. Organic Functionalization of Polyoxovanadate–Alkoxide Clusters: Improving the Solubility of Multimetallic Charge Carriers for Nonaqueous Redox Flow Batteries. ChemSusChem 2018, 11, 4139–4149. 10.1002/cssc.201802029. [DOI] [PubMed] [Google Scholar]
  51. Swinehart D. F. The Beer-Lambert Law. J. Chem. Educ. 1962, 39, 333–335. 10.1021/ed039p333. [DOI] [Google Scholar]
  52. Scrucca L. GA: A Package for Genetic Algorithms in R. J. Stat. Software 2013, 53, 1–37. 10.18637/jss.v053.i04. [DOI] [Google Scholar]
  53. Scrucca L. On Some Extensions to GA Package: Hybrid Optimisation, Parallelisation and Islands Evolution. R J. 2017, 9, 187–206. 10.32614/rj-2017-008. [DOI] [Google Scholar]
  54. Efron B. Bootstrap Methods: Another Look at the Jackknife. Ann. Stat. 1979, 7, 1–26. 10.1214/aos/1176344552. [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

ao1c05133_si_001.pdf (1.4MB, pdf)

Articles from ACS Omega are provided here courtesy of American Chemical Society

RESOURCES