Viologen-Decorated TEMPO for Neutral Aqueous Organic Redox Flow Batteries

A novel electroactive organic molecule, viz., 1-(1-oxyl-2,2,6,6-tetramethylpiperidin-4-yl)-1′-(3-(trimethylammonio)propyl)4,4′-bipyridinium trichloride ((TPABPy)Cl3), is synthesized by decorating 2,2,6,6-tetramethylpiperidin-1-oxyl (TEMPO) with viologen, which is used as the positive electrolyte in neutral aqueous redox flow battery (ARFB). Extensive characterizations are performed to investigate the composition/structure and the electrochemical behavior, revealing the favorable effect of introducing the cationic viologen group on the electroactive TEMPO. Salient findings are as follows. First, the redox potential is elevated from +0.745V for TEMPO to +0.967V for decorated TEMPO, favoring its use as the positive electrolyte. Such an elevation originates from the electron-withdrawing effect of the viologen unit, as evidenced by the nuclear magnetic resonance and single crystal structure analysis. Second, linear sweep voltammetry reveals that the diffusion coefficient is 2:97 × 10−6 cm2 s−1, and the rate constant of the one-electron transfer process is 7:50 × 10−2 cm s−1. The two values are sufficiently high as to ensure low concentration and kinetics polarization losses during the battery operation. Third, the permeability through anion-exchange membrane is as low as 1:80 × 10−11 cm2 s−1. It is understandable as the positive-charged viologen unit prevents the molecule from permeating through the anion exchange membrane by the Donnan effect. Fourth, the ionic nature features a decent conductivity and thus eliminates the use of additional supporting electrolyte. Finally, a flow battery is operated with 1.50M (TPABPy)Cl3 as the positive electrolyte, which affords a high energy density of 19.0WhL -1 and a stable cycling performance with capacity retention of 99.98% per cycle.


Introduction
Wide application of clean energy (e.g., solar and wind energy) in near future needs urgent development of electrochemical energy storage technologies [1]. Redox flow battery (ARFB) is one of the most important large-scale technologies, and all-vanadium redox flow battery (VRFB) has proved itself as one option [2]. The current challenge is the high cost of the electrolyte due to the resource scarcity of vanadium. Considerable efforts have been then devoted to developing alternative electrolytes [3]. Electroactive organic merit of element abundance, low cost, and flexible molecular control over the electrochemical features, which have been regarded as key to develop next-generation RFBs [4].
2,2,6,6-Tetramethylpiperidin-1-oxyl (TEMPO) shows reversible one-electron storage and high redox potential in organic solvent, which has been applied as the positive electrolyte in nonaqueous RFB since 2011 [5]. Hydrophilic functionalization of the molecule makes it possible be used in an aqueous redox flow battery (ARFB). For example, Liu found that the hydroxyl-functionalized TEMPO 4-hydroxy-2,2,6,6tetramethylpiperidin-1-oxyl (4-HO-TEMPO) showed a high aqueous solubility of 2.1 M [6]. Then, they developed an ARFB by using 0.5 M methyl viologen(-)/4-HO-TEMPO(+) for demonstration. Quaternization has been proved as an effective strategy to increase the solubility and redox potential. Schubert developed N,N,N-2,2,6,6-heptamethylpiperidinyl oxy-4-ammonium chloride (TEMPTMA) with a high redox potential of 0.99 V and yielded a high energy density of 25.2 Wh L -1 in an ARFB [7]. It is argued that its high potential might be electrochemically unstable in aqueous solution [8]. Yang synthesized a more stable molecule, viz., 4-[3-(trimethylammonio)propoxy]-2,2,6,6-tetramethylpiperidine-1oxyl chloride (TMAP-TEMPO), which showed a potential of 0.81 V and an energy density of 12.0 Wh L -1 in an ARFB [9]. Besides, the same author claimed that the neutral TEMPO molecule may easily penetrate the separator membrane from the positive to negative side, which threatens the battery stability during the operation. The quaternary ammonium cation can effectively prevent the molecule against its crossover through the anion exchange membrane (AEM) by the Donnan effect and thus improve the battery stability. Literature survey reveals that the practical application of TEMPO as the positive electrolyte needs more work on molecule engineering to increase the aqueous solubility, elevate the redox potential, and prevent its crossover.

Material Characterization.
Crystal data of (TPABPy)PF 6 was collected by a SuperNova diffractometer with Atlas S2 detector using a MoKα microfocus source. 1 H NMR spectra were obtained on an AVANCE III HD 500 MHz (Bruker, Germany). UV-vis spectra were measured by a Shimadzu UV-2600 spectrometer.
where P is the permeability (cm 2 s -1 ), C t is the concentration (mol L -1 ) of active species at time t in the receiving side, V 0 is the volume of the solution in either compartment, A is the effective membrane area (0.785 cm 2 ), l is the thickness of the membrane (125 μm), t is the the elapsed time (s), and C 0 is the concentration of electrolyte in the donating compartment (0.10 mol L -1 ).

Electrochemical Measurements.
Linear sweep voltammetry and cyclic voltammetry were carried out on a CHI 630E electrochemical workstation. The cyclic voltammograms (CVs) were collected at a scan rate of 100.0 mV s -1 on a standard three-electrode system (the working electrode: a glassy carbon disk (diameter: 3 mm); the reference electrode: an Ag/AgCl electrode; the counterelectrode: a platinum wire). Linear sweep voltammograms (LSVs) were collected by a rotating disk electrode (RDE, diameter: 5 mm) at a scan rate of 5.0 mV s -1 . Koutecky-Levich equation was used to determine the diffusion coefficient and the kinetic parameters: where i is the current (A) at a specific overpotential, i k is the kinetic current (A), ω is the rotating rate (rad s -1 ), C 0 is the bulk concentration of (TPABPy)Cl 3   Step 1 Step 2 Step 3 Step 4 Scheme 2: Synthesis protocol of (TPABPy)Cl 3 . 3 Energy Material Advances 2.6. Flow Battery Test. A laboratory-scale flow battery was assembled as follows. Graphite felt electrodes (3 cm × 3 cm in size, 4 mm thick) were placed beside the titanium current collectors, and a Selemion AMV anion-exchange membrane was used as the separator. A peristaltic pump (Longer BT100-2 J) was used to circulate the electrolytes through the cell at a flow rate of 60.0 mL min -1 . A Neware BTS 3000 battery testing system was used to evaluate the battery at 25°C in a 99.999% argon-filled glove box (Universal 2440/750/900). Cycling experiment was performed by galvanostatic cycling with 1.5 and 0.5 V cut offs.
The battery performance was evaluated by two concentrations of (TPABPy)Cl 3

Results and Discussion
3.1. Structure and Composition of TEMPO. As illustrated in Scheme 2, (TPABPy)Cl 3 is synthesized in a facile protocol, which includes the Zincke reaction [19], an S N 2 reaction and ion exchange step. And the purification of each step can be performed by filtration without use of   Energy Material Advances chromatography. The structure and composition of the product are investigated by 1 H-NMR (Figures S1-S2) and single crystal structure analysis ( Figure S3, Table S1). The electronic effect of viologen on TEMPO is investigated by 1 H NMR (Figure 1(a), Figure S4). Figure 1(a) reveals a remarkable chemical shift to low field for the protons of methyl groups in the TEMPO ring, which illustrates an electron-withdrawing effect of viologen. This singlet peak splits into two singlet ones due to the nuclear overhauser effect (NOE, Figure S4). X-ray single crystallographic analysis confirms that H-H distance between the protons at the 4-postion and a-CH 3 in the TEMPO ring lies within internuclear distances for NOE (<5 Å) (see the red circle in Figure 1(b)).
The energy level of the lowest unoccupied molecular orbital (LUMO) and the highest occupied molecular orbital (HOMO) of the oxidized state is calculated at the B3LYP/6-31G * level. Figure 2 shows that the LUMO of (TPABPy) 4+ (−4.572 eV) is lower than that of (TEMPO) + (−4.293 eV), indicating that (TPABPy) 4+ has a higher electron affinity than does (TEMPO) + .

Electrochemical Performance.
The redox potential is evaluated by cyclic voltammetry (CV). Figure 3 reveals that the formal potential of (TPABPy)Cl 3 (+0.967 V, vs. SHE) is 0.222 V higher than that of TEMPO (+0.745 V, vs. SHE). The elevation of the redox potential can be attributed to the remarkable electron-withdrawing effect of the viologen unit, as evidenced by both NMR and the energy level analysis (vide supra). Both the diffusion coefficient and the rate constant are derived from linear sweep voltammetry (LSV, Figures S5-S8). The diffusion coefficient of (TPABPy)Cl 3 is 2:97 × 10 −6 cm 2 s −1 , which is slightly lower than that of TEMPO (see Table 1). The decrease is understandable 5 Energy Material Advances considering the higher molecular weight after decorating viologen. The rate constant of the one-electron oxidation reaction of (TPABPy)Cl 3 is 7:50 × 10 −2 cm s −1 , which is higher than that of TEMPO (6:62 × 10 −3 cm s −1 ) and TMAP-TEMPO [9] (1:02 × 10 −2 cm s −1 ). This result illustrates that the viologen unit favors the electron transfer of TEMPO during the redox reaction. The above features of the (TPABPy)Cl 3 make it as a promising positive electrolyte for ARFBs.

Flow Battery Performance.
As demonstrated, a battery is operated with 1.5 M (TPABPy)Cl 3 as the positive electrolyte and 1.2 M BTMAP-Vi as the negative electrolyte. Notably, no external supporting electrolyte is added in either side due to their high conductivity. Figure 3 reveals that the standard voltage of this battery is up to 1.30 V. Figure 4(a) displays the rate performance with operating current from 20 to 80 mA cm -2 . It is seen that the battery yields a high capacity of 37.6 Ah L -1 at 20 mA cm -2 , and the capacity shows a mild decrease with increasing the current density. Figure 4(b) shows that the battery delivers a nearly 100% coulombic efficiency (CE) at each current density. The energy efficiency (EE) and voltage efficiency (VE) are found to be sufficiently high, e.g., 70.6% EE and 71.1% VE achieved at 60 mA cm -2 . The high efficiency can be attributed to the high electrochemical reversibility of the two redox couples. Figure 4(c) shows the 100-cycle test result at 60 mA cm −2 to evaluate the stability. This battery provides a high capacity retention of 99.98% per cycle, demonstrating its superior stability. It is noted that this battery represents the most stable one in the reported viologen(-)/TEMPO(+) ARFBs without using any additional supporting electrolyte ( Figure 5) [6,9,[20][21][22][23][24]. Postcell analysis is conducted by using CV and 1 H NMR to investigate the superior stability. As shown in Figure 6(a), the redox signal of (TPABPy)Cl 3 is not seen in the negative electrolyte, suggesting the crossover of (TPABPy)Cl 3 is negligible. In addition, Figure 6(b) only shows the proton resonance signals of the BTMAP-Vi. These results strongly verify the low crossover of the electroactive species in both sides, which contributes to the cycling stability of the battery.

Conclusions
In this work, viologen-decorated TEMPO ((TPABPy)Cl 3 ) is developed as the positive electrolyte in neutral aqueous organic redox flow battery. The introduction of viologen unit is proved effective in increasing the TEMPO's aqueous solubility, elevating the redox potential and preventing its crossover. As a result, by pairing BTMAP-Vi as the negative electrolyte, an assembled flow battery using 1.5 M (TPAB-Py)Cl 3 in the positive side without any supporting electrolyte affords a high energy density of 19.0 Wh L -1 and a stable  Energy Material Advances cycling performance with capacity retention of 99.98% per cycle. This work highlights the rational design of positive electrolyte for neutral ARFB, which will pave the road to develop energy-dense, stable, and lost-cost ARFBs.

Data Availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.