BIOBASED AQUEOUS ORGANIC-BASED ELECTROLYTE IN AQUEOUS ORGANIC REDOX FLOW BATTERY

Abstract

A redox flow battery includes at least one aqueous electrolyte including the fermented compound

##STR00001##

and/or an ion, and/or a salt and/or a reduced form of the anthraquinone member of compound (I). X.sub.1-X.sub.8 are independently selected from a hydrogen atom, a halogen atom, an ether group of formula O-A, a linear, cyclic or branched, saturated or unsaturated, optionally substituted, hydrocarbon group including 1-10 carbon atoms, a OH group, a R.sub.1 group and a O-A-R.sub.1 group. A represents a linear, cyclic or branched, saturated or unsaturated, optionally substituted, hydrocarbon group including from 1-10 carbon atoms. A represents a linear, cyclic or branched, saturated or unsaturated, optionally substituted, hydrocarbon group including 1-10 carbon atoms. R.sub.1 represents COOH, SO.sub.3H, or a salt thereof. One to three of X.sub.1-X.sub.8 is OH. Exactly one of X.sub.1-X.sub.8 is O-A-R.sub.1. A method for generating electricity with such compounds is also described.

Claims

1. A redox flow battery comprising at least one aqueous electrolyte comprising a fermented compound of formula (I) ##STR00014## and/or an ion of compound (I), and/or a salt of compound (I), and/or a reduced form of the anthraquinone member of compound (I), wherein: X.sub.1, X.sub.2, X.sub.3, X.sub.4, X.sub.5, X.sub.6, X.sub.7 and X.sub.8 are independently selected from the group consisting of a hydrogen atom, a halogen atom, an ether group of formula O-A, a linear, cyclic or branched, saturated or unsaturated, optionally substituted, hydrocarbon group comprising from 1 to 10 carbon atoms, a OH group, a R.sub.1 group and a O-A-R.sub.1 group, A represents a linear, cyclic or branched, saturated or unsaturated, optionally substituted, hydrocarbon group comprising from 1 to 10 carbon atoms; A represents a linear, cyclic or branched, saturated or unsaturated, optionally substituted, hydrocarbon group comprising from 1 to 10 carbon atoms; R.sub.1 represents COOH, SO.sub.3H, or a salt thereof; wherein one to three of X.sub.1, X.sub.2, X.sub.3, X.sub.4, X.sub.5, X.sub.6, X.sub.7 and X.sub.8 is OH, and wherein one and only one of X.sub.1, X.sub.2, X.sub.3, X.sub.4, X.sub.5, X.sub.6, X.sub.7 and X.sub.8 is O-A-R.sub.1.

2. The redox flow battery according to claim 1, wherein one and only one of X.sub.1, X.sub.2, X.sub.3, X.sub.4, X.sub.5, X.sub.6, X.sub.7 and X.sub.8 is OH and wherein one and only one of X.sub.1, X.sub.2, X.sub.3, X.sub.4, X.sub.5, X.sub.6, X.sub.7 and X.sub.8 is O-A-R.sub.1.

3. The redox flow battery according to claim 1, wherein one and only one of X.sub.3 and X.sub.8 is OH, and one and only one of X.sub.3 and X.sub.8 is O-A-R.sub.1.

4. The redox flow battery according to claim 1, wherein X.sub.3 is OH and X.sub.8 is O-A-R.sub.1.

5. The redox flow battery according to claim 1, wherein X.sub.3 is O-A-R.sub.1 and X.sub.8 is OH.

6. The redox flow battery according to claim 1, wherein two of X.sub.3, X.sub.6, and X.sub.8 are OH, and wherein one and only one of X.sub.3, X.sub.6 and X.sub.8 is O-A-R.sub.1.

7. The redox flow battery according to claim 1, wherein three of X.sub.3, X.sub.5, X.sub.6, and X.sub.8 are OH, and wherein one and only one of X.sub.3, X.sub.5, X.sub.6 and X.sub.8 is O-A-R.sub.1.

8. The redox flow battery according to claim 1, wherein X.sub.1 represents a linear, cyclic or branched, saturated or unsaturated, optionally substituted, hydrocarbon group comprising from 1 to 10 carbon atoms.

9. The redox flow battery according to claim 1, wherein A represents (CH.sub.2).sub.n, optionally substituted, wherein n is an integer selected from 1 to 10.

10. The redox flow battery according to claim 1, wherein said O-A-R.sub.1 group represents O(CH.sub.2).sub.nCOOH or O(CH.sub.2).sub.n SO.sub.3H wherein n is from 1 to 10.

11. The redox flow battery according to claim 1, wherein R.sub.1 is selected from the group consisting of CO.sub.2H, CO.sub.2.sup.?M.sup.+, SO.sub.3H and SO.sub.3.sup.?M.sup.+, M.sup.+ being selected from the group consisting of Li.sup.+, Na.sup.+, K.sup.+ and NH.sub.4.sup.+.

12. The redox flow battery according to claim 1, wherein X.sub.2 represents a R.sub.1 group.

13. The redox flow battery according to claim 1, wherein said compound of formula (I) has the following structure: ##STR00015## wherein X.sub.1, X.sub.2, X.sub.3, and X.sub.8 are as defined in claim 1.

14. The redox flow battery according to claim 1, wherein said compound of formula (I) has the following structure: ##STR00016## wherein X.sub.1, X.sub.2 and O-A-R.sub.1 are as defined in claim 1.

15. The redox flow battery according to claim 1, wherein said compound of formula (I) has the following structure: ##STR00017## wherein O-A-R.sub.1 are as defined in claim 1.

16. The redox flow battery according to claim 1, wherein said compound of formula (I) has the following structure: ##STR00018## wherein O-A-R.sub.1 is as defined in claim 1.

17. A method for generating electricity by at least one redox flow battery, wherein the redox flow battery is defined according to claim 1.

Description

FIGURES

In the Figures:

[0119] FIG. 1 is a graph representing (squares) the charge capacity, (circles) the discharge capacity, (diamonds) the coulombic and (triangles) the energy efficiencies of a cell test involving a negolyte with compound according to the present invention

[0120] FIG. 2 represents the evolution of (dark triangles) coulombic and (dark squares) energy efficiencies according to the nominal power density during the same cycling test involving a negolyte containing compound.

[0121] FIG. 3: Cyclic voltammetry (3.sup.rd cycle) of the negolyte (black line) before the cycling test and (grey line) after cycling test, recorded at a scan rate of 100 mV.Math.s.sup.?1.

[0122] FIG. 4: .sup.1H NMR spectra of the negolyte, recorded in D.sub.2O between 0 and 10 ppm, (black line) before and (grey line) after the cycling test.

[0123] FIG. 5 is a graph representing (squares) the capacity retention, (diamonds) coulombic and (triangles) energy efficiencies of a cell test involving a negolyte with a compound according to the present invention. 500 cycles are presented.

[0124] FIG. 6 is a graph representing the capacity available during charge and during discharge (dark, dark grey and light grey) of a cell test involving a negolyte with a compound according to the present invention.

[0125] FIG. 7: Cyclic voltammetry (3.sup.rd cycle) of the negolyte (black line) before the cycling test and (grey line) after cycling test, recorded at a scan rate of 100 mV.Math.s.sup.?1.

[0126] FIG. 8: .sup.1H NMR spectra of the negolyte, recorded in D.sub.2O between 0 and 10 ppm, before (above spectra) and after (bottom spectra) the cycling test.

[0127] FIG. 9 is a graph representing (squares) the capacity retention, (diamonds) coulombic and (triangles) energy efficiencies of a cell test involving a negolyte with with a compound according to the present invention. 100 cycles are presented.

[0128] FIG. 10: Cyclic voltammetry (3.sup.rd cycle) of the negolyte (black line) before the cycling test and (grey line) after cycling test, recorded at a scan rate of 100 mV.Math.s.sup.?1.

[0129] FIG. 11: .sup.1H NMR spectra of the negolyte, recorded in D.sub.2O between 0 and 10 ppm, (above spectra) before and (bottom spectra) after the cycling test.

EXAMPLES

1. Production of Anthraquinone Molecule A and B by Fermentation:

[0130] Molecule A and B can be produced by fermentation of microorganisms expressing the necessary genes to their synthesis such as an octaketide synthase, a ketoreductase, an aromatase and a cyclase and fermented in a minimal medium supplemented with glucose and inoculated as known by the skilled man of the art. Alternatively, bioproduction of the molecules can be performed enzymatically in vitro with supplementation of precursors such as malonyl or acetyl-CoA as known by the man skilled in the art.

[0131] Production and extraction of those molecules can be performed in a recombinant E. coli as exhibited in patent demand WO2020161354A1 or in a modified Streptomyces CH999 as exhibited in patent demand WO199508548, or in Taek Son Lee et al., Biorg. Med. Chem., 2007, 15, 5207 or in vitro as described in Carreras et al., J. Am. Chem. Soc. 1996, 118, 5158-5159.

NMR Procedure

[0132] DMSO-d6 was purchased from Euriso-top (France). 2-thiohydantoin purchased from Alfa Aesar is used internal standard to assess the sample purity as well as maleic acid purchased from Acros Organics. Product and 2-thiohydantoin were dissolved in 0.5 mL of DMSO-d6. .sup.1H NMR spectra were recorded on BRUKER AC 300 P (300 MHz) spectrometer. Chemical shifts are expressed in parts per million. Data are given in the following order: value, multiplicity (s, singlet; d, doublet; t, triplet; q, quadruplet; p, quintuplet; m, multiplet), coupling constants J (given in Hertz), number of protons.

##STR00010##

Molecule A: NMR Signature and Purity

[0133] .sup.1H NMR (300 MHz, DMSO) ? 12.87 (s, 1H), 7.72 (t, J=7.9 Hz, 1H), 7.63 (dd, J=7.6, 1.3 Hz, 1H), 7.59 (s, 1H), 7.33 (dd, J=8.3, 1.3 Hz, 1H), 2.69 (s, 3H).

[0134] The purity of molecule A is estimated by quantitative .sup.1H NMR by using an internal standard, the 2-Thiohydantoin: .sup.1H NMR (300 MHz, DMSO) ? 11.64 (s, 1H), 9.83 (s, 1H), 4.04 (s, 2H). The purity is estimated at 78%.

##STR00011##

Molecule B: NMR Signature and Purity

[0135] .sup.1H NMR (300 MHz, DMSO) ? 12.87 (s, 1H), 10.99 (s, 1H), 7.61 (t, J=7.8 Hz, 1H), 7.50 (d, J=7.4 Hz, 1H), 7.32 (s, 1H), 7.21 (d, J=8.2 Hz, 1H), 6.90 (s, 1H), 2.59 (s, 3H).

[0136] The purity of molecule A is estimated by quantitative .sup.1H NMR by using an internal standard, the 2-Thiohydantoin: .sup.1H NMR (300 MHz, DMSO) ? 11.64 (s, 1H), 9.83 (s, 1H), 4.04 (s, 2H). The purity is estimated at 96%.

2. Synthesis of Molecular Derivatives Using an Organic Chemistry Process. Purity Measurement.

[0137] Potassium carbonate and potassium hydroxide (purity 85%) were purchased from VWR. Dry dimethylformamide and methyl 4-bromobutyrate were purchased from Acros Organics. Glacial acetic acid was purchased from Carlo Erba. Isopropanol and concentrated hydrochloric acid (37%) were purchased from Sigma Aldrich.

Synthesis of methyl 4-((5-hydroxy-4-methyl-9,10-dioxo-9,10-dihydroanthracen-2-yl)oxy)butanoate

[0138] ##STR00012##

[0139] 3,8-dihydroxy-1-methylanthracene-9,10-dione (5 g; 19.67 mmol; 1 eq) were dissolved in 150 mL of dry dimethylformamide with potassium carbonate (2.99 g; 21.63 mmol; 1.1 eq). The mixture is stirred at 50? C. and a solution of methyl 4-bromobutyrate (3.56 g; 19.67 mmol; 1 eq) in 50 mL of dry dimethylformamide is added dropwise over a period of 30 minutes. After stirring 44 hours at 50? C., the mixture is poured on 125 mL of a solution of hydrochloric acid at 4%. The precipitate is isolated by filtration on a frit glass and washed with 100 mL of ultrapure water. The product is dried under vacuum for 6 hours at 40? C., leading to the desired compound an ocher powder with a yield of 83% (purity was measured by .sup.1H NMR at 100% by using maleic acid as an internal standard).

[0140] .sup.1H NMR (300 MHz, DMSO) ? 12.82 (s, 1H), 7.70 (t, J=7.9 Hz, 1H), 7.59 (d, J=7.4 Hz, 1H), 7.42 (d, J=2.7 Hz, 1H), 7.30 (d, J=8.3 Hz, 1H), 7.22-7.10 (m, 1H), 4.15 (t, J=6.4 Hz, 2H), 2.68 (s, 3H), 2.02 (p, J=6.8 Hz, 2H).

[0141] Internal standard (maleic acid): .sup.1H NMR (300 MHz, DMSO) ? 6.27 (s, 2H).

Synthesis of 4-((5-hydroxy-4-methyl-9,10-dioxo-9,10-dihydroanthracen-2-yl)oxy)butanoic acid (compound B-M3CH)

[0142] ##STR00013##

[0143] Methyl 4-((5-hydroxy-4-methyl-9,10-dioxo-9,10-dihydroanthracen-2-yl)oxy)butanoate, (5.8 g; 16.37 mmol; 1 eq) is suspended in 250 mL of ultrapure water and 130 mL of isopropanol. Potassium hydroxide is added (1.89 g; 28.64 mmol; 1.75 eq) and the dark red solution is stirred at 60? C. for 24 hours. The solution is poured on 700 mL of ice and 17 mL of acetic acid is added to induce precipitation of the product which is collected by filtration on a glass frit. The product is washed with 50 mL of ultrapure water and dried under vacuum for 14 hours at 40? C. The desired product is isolated as an ocher powder with a yield of 91% (purity was measured by .sup.1H NMR at 95% by using maleic acid as an internal standard).

[0144] .sup.1H NMR (300 MHz, DMSO) ? 12.82 (s, 1H), 7.71 (t, J=7.9 Hz, 1H), 7.61 (dd, J=7.6, 1.3 Hz, 1H), 7.45 (d, J=2.7 Hz, 1H), 7.31 (dd, J=8.2, 1.2 Hz, 1H), 7.21 (d, J=2.7 Hz, 1H), 4.17 (t, J=6.4 Hz, 2H), 2.70 (s, 3H), 2.43 (t, J=7.3 Hz, 2H), 2.00 (p, J=6.9 Hz, 2H).

[0145] Internal standard (maleic acid): .sup.1H NMR (300 MHz, DMSO) ? 6.27 (s, 2H).

3. Water Solubilities

Experimental Procedure

[0146] The solubility of the compounds was evaluated in aqueous solution at various pH by shake-flask method. An aqueous solution (containing either KOH 1 M, NaOH 1 M, or a mixture of KOH 0.5 M and NaOH 0.5 M) is added over 100 mg of compound by 20 ?L portions until complete solubilization. We obtain an approximate range of solubility values calculated as following:

Smax?(m/M)/V; m=mass of evaluated compound(g);M=molar mass of evaluated compound(g.Math.mol.sup.?1);V=added volume(L).

Results

[0147]

TABLE-US-00001 TABLE 1 solubility estimated of compounds A, B and of one compound of formula (I) (compound B-M3CH) using shake-flask method KOH 0.5M + Compound (I) KOH (1M) NaOH (1M) NaOH 0.5M A 0.48 0.32 0.49 B 0.58 0.5 0.63 B-M3CH 0.15 0.15 0.43

[0148] Compounds A and B are readily soluble (see table 1) in aqueous alkaline environment (>0.3 M) and present, surprisingly, an even higher solubility when mixing hydroxide salts holding potassium and sodium cations.

[0149] Compound B-M3CH of formula (I) exhibits a rather low solubility, compared to compounds A and B, in ionic environment containing one type of alkaline cation (?0.15 M). When in presence with a mix of cations (sodium and potassium cations in this example), the water solubility of this compound is increased by a factor close to 3 (?0.43 M). This result highlights the major influence of the cationic environment on the water solubility of such compound.

4. Electrochemical Properties at Diluted State

Experimental Procedure

[0150] Cyclic voltammetry experiments were performed at a 100 mV/s scan rate for compounds A, B and B-M3CH in a solution at pH 14 (KOH 1 M).

[0151] The electrochemical experiments were conducted on an electrochemical working station (BioLogic Science Instruments VSP) at 25? C. using a three-electrodes electrochemical cell, where a Pt wire was employed at the counter electrode (CE), an Ag/AgCl/KCl saturated electrode immerged in a 1 M KOH solution protected from the cell by a porous frit glass served as the reference electrode (RE) and a rotating disc electrode (RDE) mounted with a glassy carbon (GC, 7.069 mm.sup.2) disk was used as the working electrode (WE), respectively. The potential was reported relative to the normal hydrogen electrode (NHE), which was converted from the Ag/AgCl reference electrode (0.230 V versus NHE). The ohmic drop between the WE and RE is determined by current interrupt experiments and leads to a value close to 0.7069 ?.Math.cm.sup.2. All cyclic voltammetry experiments were conducted by alleviating most of this resistance (85%) using the potentiostat manual ohmic drop compensation. Before each electrochemical measurement, the GC disk electrode was polished using SiC P4000 foil.

[0152] Each recorded cyclic voltammetry of the reduction and the oxidation process of a compound at a given scan rate is analyzed by measuring i.sub.p,a=anodic peak current; i.sub.p,c=cathodic peak current; v=scan rate; E.sub.p,a=anodic peak potential; E.sub.p,c=cathodic peak potential; ?E.sub.p=E.sub.p,a?E.sub.p,c; the apparent standard redox potential E.sup.0=(E.sub.p,a?E.sub.p,c)/2.

[0153] Linear sweep voltammetry (LSV) using the RDE at various rotation speed (400 up to 1 000 RPM) were performed to determine the averaged half-wave potential E.sub.1/2.

[0154] The oxidation (DO) and the reduction (DR) diffusion coefficients were calculated using Randles-Sevcik (RS) equation i.sub.p=0.4463.n.sup.3/2.F.sup.3/2.A.C.v.sup.1/2.D.sup.1/2.R.sup.?1/2.T.sup.?1/2=Cte.v.sup.1/2.D.sup.1/2, n=number of electrons exchanged during the redox process; F=faradaic number=96485 C.Math.mol.sup.?1; A=electrode active surface; C=redox active molecule concentration; R=gas constant=8.314 J.Math.K.sup.?1.Math.mol.sup.?1; T=bulk temperature=299 K. The experiments are conducted by cyclic voltammetry at various scan rates ranging from 10 mV/s up to 20 V/s.

[0155] A graphic of |i.sub.p,a| and |i.sub.p,c|=f(Cte.v.sup.1/2) is plotted, if the redox process is reversible and diffusion controlled, the peak current should be proportional to the square root of the scan rate. The diffusion coefficients (oxidation and reduction) are then calculated by taking the square of the regression lines slopes. Diffusion coefficients (oxidation and reduction processes) are evaluated by plotting the peak current versus a constant multiplied by the square root of the scan rate. The plots regression line gives a straight line with perfect correlation coefficient (100%) confirming that the redox process is reversible and diffusion controlled. The square of each slope gives the associated coefficient diffusion in cm.sup.2.Math.s.sup.?1.

[0156] The apparent heterogeneous rate constant during reduction (k.sup.0.sub.O) process was calculated using Gileadi's equation (C. Russel, W. Jaenicke, Electrochimica Acta, Vol 27, No. 12, 1745-1750, 1982) log k.sup.0=?0.48.?+0.52+log[(n.F.?.v.sub.c.D)/(2.303.R.T)].sup.1/2; v.sub.c is the critical scan rate, it is calculated by plotting the peak potentials versus the logarithm of the scan rates, and is more precise when a very wide range of scan rates is used, which is the case here (0.010 up to 20 V.Math.s.sup.?1), it is determined at the x-axis of the point of intersection of the regression lines drawn at low and high scan rates; a is the transfer coefficient that is supposed to be close to 0.5 for all presented compounds.

[0157] Peak-to-peak separation ?E.sub.p values were determined from the third cycle of each cyclic voltammogram and are gathered in the following table 2.

TABLE-US-00002 TABLE 2 Peak-to-peak separation ?E.sub.p of compounds A, B and of one compound of formula (I) at pH 14 (compound B-M3CH) at a concentration of 5.10.sup.?3 M at a scan rate of 100 mV .Math. s.sup.?1 Compounds ?E.sub.p (mV) at pH 14 A 56 B 40 B-M3CH 30

[0158] The standard redox process reversibility evaluation is performed in diluted state, at 5 mM of concentration, at pH 14 buffered with 1 M of KOH. Alan J. Bard et al. (A. J. Bard and L. R. Faulkner, Electrochemical Methods, Wiley, New York, 2.sup.nd edition, 2001) presents the 5 criteria defining the reversibility of a redox process in the handbook of electrochemical methods: [0159] i) the peak current absolute values must be proportional to the square root of the scan rate; [0160] ii) the peak potentials must be independent at all scan rates; [0161] iii) the potential peak difference must be equal to 59/n if no protons are involved, n=number of electrons involved in the redox process (2 electrons in this case); [0162] iv) the ratio of peak current absolute values must be equal to the unity; [0163] v) the apparent standard redox potential (E.sup.0) must be equal to the half-wave potential (E.sub.1/2).

[0164] Main electrochemical properties of compound of formula (I) are gathered in table 3.

TABLE-US-00003 TABLE 3 Electrochemical parameters of compounds A and B and one compound of formula (I) (compound B-M3CH) E.sup.0 E.sub.1/2 D.sub.O ? 10.sup.?6 k.sup.0.sub.O ? 10.sup.?2 Compound (I) (V vs NHE) (V vs NHE) (cm.sup.2/s) (cm/s) A ?0.59 ?0.59 3.65 0.78 B ?0.57 ?0.57 2.99 0.80 B-M3CH ?0.49 ?0.49 4.58 2.47

[0165] Redox potentials of compound of A and B are are close to ?0.6 V vs NHE which allows to reach an open-circuit-voltage (ocv) of about 1.1 V at 50% state-of-charge (SOC) in a cell involving standard ferricyanide/ferrocyanide redox couple (E.sup.0=0.5 V vs NHE) in the posolyte.

[0166] Redox potentials of compounds of formula (I), i.e. compound B-M3CH, are close to 0.50 V vs NHE which allows to reach an ocv of about 1.0 V at 50% SOC in a cell involving standard ferricyanide/ferrocyanide redox couple in the posolyte.

[0167] The redox process is fast and reversible for a compound of formula (I). This result is important as it discloses the ability of the redox active material to undergo multiple charge and discharge processes without changes of electrochemical properties and performances.

[0168] The diffusion coefficients values vary from 3 to 5?10.sup.?6 cm.sup.2/s which is typical of redox active materials used in aqueous redox flow battery (ARFB) negolytes in aqueous medium. Inversely, the electron transfer kinetics (.sub.O stands for oxidant, .sup.0 stands for standard, .sup.0 stands for standard apparent) obtained with compounds of formula (I) (>10.sup.?2 cm/s) are faster than most redox active materials used in ARFB negolytes (see table 4).

TABLE-US-00004 TABLE 4 Diffusion coefficients and electron transfer kinetics of various redox-active molecules used in negolytes in ARFBs Redox active D.sub.O ? 10.sup.?6 k.sup.0.sub.O ? 10.sup.?2 material in negolytes (cm.sup.2/s) (cm/s) DHBQ 3.66 0.21 DHAQ 4.8 0.70 DBEAQ 1.58 0.70 DPPEAQ 1.37 Flavine FMN-Na 1.3 0.53 AQ-2,7-DS 3.8 0.72 BTMAP-Vi 3.3 2.2 V.sup.3+/V.sup.2+ 4 0.0017

5. Cycling Tests in AORFB Environment

[0169] The use of a very specific redox flow battery system is essential to allow the energy harvesting of these bio-fermented active molecules. The requisite know-how relates to a set of multiple technological bricks that have been developed together to fit the specifications imposed by organic-based electrolytes in AORFBs.

[0170] The electrolyte composition (osmotic flux management, pH buffering, ionic conductivity optimization . . . ); the operational conditions (cycling conditions, in-situ analytical procedures . . . ); the materials choice (pretreatments, electrode material, membrane nature, the interaction with the chemicals and the environment . . . ); the process (hydraulic circuit, impact of the system layout on the performance, the tank design . . . ); the system control (process parameter monitoring, BMS . . . ); the stack design (architecture, materials suitability, electrode design, shunt current management . . . ) represent a non-exhaustive list of specific knowledges that, in synergy, allow the development of a redox flow battery that is compatible with organic-based electrolytes.

[0171] In addition, specific know-how in the AORFB field allows to evaluate such electrolyte composed with a bio-fermented organic active material under near-real operational conditions. Namely, the cycling tests are performed during a relevant period (above 10 days of constant cycling which is the average RFB cycling period in the state-of-the-art); using very low (relevant with the system's scale) electronic excess in the non-limiting capacity compartment; using very low volume excess (relevant with the system's scale) in the non-limiting capacity compartment; using low flow rates (relevant with the system's scale); reaching more than 4 hours of energy storage in discharge; using power constant mode during cycling rather than galvanostatic mode.

Experimental Procedure

[0172] The electrochemical diagnosis of the compounds A, B, and B-M3CH that was performed in diluted state in a standard three-electrode electrochemical cell was followed by an evaluation in concentrated state (?0.1 M) where most secondary reactions linked to the redox active degradation are likely to occur.

[0173] General procedure: Cell tests were carried out on a BCS-815 battery cycler (Biologic) at room temperature. The redox flow battery cell was composed of an Aquivion? membrane selected from E98-09S Solvay (80 microns thickness) or E98-05 Solvay (50 ?m thickness) serving as separator sandwiched between two GFD porous graphite felts (SGL Carbon group, GFD 4.6 EA), 4.6 mm thick with an active area of 5 or 25 cm.sup.2. The electrodes were inserted into 3 mm PVC frames surrounded by EPDM gaskets. The current collectors were made of graphite composite material provided by SGL and the assembly was compressed by two PVC end-plates. The electrolytes were pumped within the cells through remote-control diaphragm liquid dosing pumps (KNF) at a flow rate of 100 mL/min. The tanks were made of polyethylene connected to polyurethane tubing (4 mm inner diameter). Before each test, the cell materials (membrane and electrodes) were carefully rinsed by using 100 mL of distilled water in each tank that circulates into the cell for one hour. The cell, the process and the tanks were air purged and another rinsing step were performed for one hour before air-purging the whole system. The negolyte, which charged (reduced) form is sensitive to oxygen, was protected by an argon blanket atmosphere by bubbling the gas directly into the negolyte that was kept under circulation for 10 minutes.

[0174] Capacity utilization was calculated by determining the ratio of the discharge capacity to the theoretical capacity available in the electrolytes. Coulombic efficiency was calculated by determining the ratio of the discharge capacity to the charge capacity. Energy efficiency was calculated by determining the ratio of the discharge energy to the charge energy. The energy density was calculated by determining the ratio of the discharge energy to the total volume of posolyte (positive electrolyte) and negolyte (negative electrolyte).

[0175] Under galvanostatic mode, when not specified, the standard operating current density used was 40 mA/cm.sup.2.

[0176] Under power constant mode, when not specified, the standard operating current density used was 40 mW/cm.sup.2 and a constant voltage step was used at the end of each charge at a specific upper voltage limit.

Compound A

Cycling Test

[0177] This cell test was performed using materials described above and with the following parameters: [0178] Active area: 25 cm.sup.2 [0179] Membrane: Aquivion? with 80 microns thickness [0180] Flowrate: 100 mL/min [0181] Negative pole: 40 mL of negolyte comprising 0.24 M of compound A, 0.9 M of KOH and 0.3 M NaOH [0182] Positive pole: 42 mL of posolyte comprising 0.6 M potassium ferrocyanide and 0.3 M NaOH [0183] Averaged pH: 13.5 [0184] Cycling test: galvanostatic mode or power static mode [0185] Voltage limits: between 0.7 and 1.4 V under galvanostatic mode. Under power static mode, the lower voltage limit was 0.8 V. The upper voltage limit was selected between 1.2 and 1.23 V which, in each case, was followed by a CV step at the same voltage until 0.25 W limit is reached.

[0186] FIG. 1 presents the performances of the 350 first cycles. FIG. 1 is a graph representing (squares) the charge capacity, (circles) the discharge capacity, (diamonds) the coulombic and (triangles) the energy efficiencies of a cell test involving a negolyte with compound A.

[0187] Various cycling conditions were used all along this battery test. First, a galvanostatic mode (40 mA/cm.sup.2) was used for 58 cycles. Then, a power static mode was used (40 mW/cm.sup.2) with a cutoff upper voltage of 1.2V during 44 cycles. In order to harvest more capacity, a cutoff voltage was then set at 1.23V and various power densities were tested i.e. 40 mW/cm.sup.2 for 21 cycles; 60 mW/cm.sup.2 for 22 cycles; 80 mW/cm.sup.2 for 94 cycles and, finally, 100 mW/cm.sup.2 for 131 cycles.

[0188] All available capacity (capacity utilization) was reached (100%) at the beginning of the cycling test (which makes sense as the electrochemical kinetics of redox active materials are fast and the important flowrate decrease mass transport limitations) whereas a steady capacity loss (between 0.042% loss and 0.123% capacity loss per cycle) was observed during each cycling step, whatever the current and power densities. In addition, rather low coulombic efficiencies (<99%) were observed during cycles under power static mode while high (>70%) energy efficiencies were observed even at very high nominal power densities (100 mW/cm.sup.2), showing that very high nominal power densities can be used with such systems, see FIG. 2.

[0189] FIG. 2 represents the evolution of (dark triangles) coulombic and (dark squares) energy efficiencies according to the nominal power density during the same cycling test involving a negolyte containing compound A.

[0190] These 2 observations (important capacity loss/cycle and low coulombic efficiencies) indicate that there is likely an unwanted secondary reaction that occurred during the test (we excluded osmosis, cross-over, electrolysis or air tightness failure problematics).

[0191] Furthermore, at 50% SOC, where the system does not suffer from charge-transfer or mass-transport limitations, a linar polarization curve (LPC) was regularly performed during the cycling test in order to track the active surface resistance (ASR) of the cell. The LPC was always realized in charge and discharge using a ramp of 100 A/min and ?100 A/min from 0 to 2 A and from 0 to ?2 A in order to evaluate the active surface resistance of the cell.

[0192] The active surface resistance was quite stable but fluctuated around 2.0-2.3 ?.Math.cm.sup.2, which was consistent with an electrochemical cell comprising a Nafion-type membrane with a thickness of 80 ?m which diffuses a majority of potassium cations (alkaline medium).

Post-Cycling Analyses

[0193] A preliminary electrochemical test was performed on the negolyte before and after the cycling test (FIG. 3).

[0194] The typical electrochemical signature of compound A was registered before the cycling test. The redox signal was centered at ?0.83 V vs Ag/AgCl (or ?0.60 V vs NHE) with a max current of roughly ?4 mA during reduction step and a max current of roughly 3 mA during reoxidation step. After cycling, the redox signature corresponding to the compound A is decreased in current intensity with a factor x10. The osmosis only can't explain this major signature decrease. Also, an oxidation occurs at ?0.57 V vs Ag/AgCl that was not originally present in the negolyte.

[0195] Further analyses were performed using .sup.1H NMR spectroscopy, see FIG. 4.

[0196] The protons signature is very different. The aromatic protons situated between 7.4 and 7.7 ppm, representative of the anthraquinone structure, are partly present after cycling but additional peaks appear between 6.4 and 7.4 ppm. The protons attached to the methyl group presents at 3.31 ppm before cycling, disappears after cycling while 2 singlets appear at 3.70 and 3.76 ppm. These very different protons signature suggest that the compound A was degraded into multiple compounds during the cycling test.

[0197] Considering these analyses, the compound A allows to reach high nominal power densities (100 mW/cm.sup.2) with decent energy densities (3.7 Wh/L) but a low chemical stability was highlighted by multiple indicators and implies that this compound is not adapted to the application in such ionic environment using such cycling conditions.

Compound B

Cycling Test

[0198] This cell test was performed using materials described above and with the following parameters: [0199] Active area: 5 cm.sup.2 [0200] Membrane: Aquivion? with 80 microns thickness [0201] Flowrate: 100 mL/min [0202] Negative pole: 20 mL of negolyte comprising 0.18 M of compound B and 1.4 M of KOH [0203] Positive pole: 50 mL of posolyte comprising 0.2 M potassium ferrocyanide and 1 M KOH [0204] Averaged pH: 14.0 [0205] Cycling test: galvanostatic mode [0206] Voltage limits: The upper cutoff voltage was fixed at 1.31 V (current density fixed at 40 mA/cm.sup.2) while the lower cutoff voltage depended on the nominal current density i.e. 0.75 V or 0.80 V when running at 40-80 mA/cm.sup.2 and 0.70 V when increasing the current density up to 120 mA/cm.sup.2.

[0207] FIG. 5 presents 500 typical cycles of this battery test.

[0208] FIG. 5 is a graph representing (squares) the capacity retention, (diamonds) coulombic and (triangles) energy efficiencies of a cell test involving a negolyte with compound B. 500 cycles are represented, the current densities in discharge and charge were fixed at 40 mA/cm.sup.2 with 0.80 to 1.40 V cutoff voltages.

[0209] The experimental capacity achieved was 78% at the beginning of the cycling test whereas a steady capacity loss of 0.025% per cycle was observed. High average coulombic efficiencies were recorded during these 500 cycles (?99.82%) with high (?81.47%) average energy efficiencies.

[0210] The battery was then further operated and different current densities were tested, see FIG. 6.

[0211] FIG. 6 is a graph representing the capacity available during charge at 40 mA/cm.sup.2 and during discharge (dark) at 40, (dark grey) at 80 and (light grey) at 120 mA/cm.sup.2 of a cell test involving a negolyte with compound B

[0212] Air sealing problematics were encountered during this battery test (due to very low electrolyte volumes in small tanks that are difficult to perfectly seal), which lowered the available capacity (we excluded cross-over, electrolysis or irreversible chemical degradation problematics) below 40%. Still, despite this problematic, we can observe that such battery can be cycled using high nominal current densities.

[0213] Furthermore, at 50% SOC, the ASR was tracked during the cycling test using standard procedure, The active surface resistance was quite stable but fluctuated around 2.0-2.2 ?.Math.cm.sup.2, which was consistent with an electrochemical cell comprising a Nafion-type membrane with a thickness of 80 ?m which diffuses a majority of potassium cations (alkaline medium).

Post-Cycling Analyses

[0214] A preliminary electrochemical test was performed on the negolyte before and after thousands of cycles and a few months of cycling test (FIG. 7).

[0215] The typical electrochemical signature of compound B was registered before the cycling test. The redox signal was centered at ?0.80 V vs Ag/AgCl (or ?0.57 V vs NHE) with a max current of roughly ?4 mA during reduction step and a max current of roughly 2.8 mA during reoxidation step. After cycling, the redox signature corresponding to the compound B is a bit decreased in current intensity, about 30-35%, which corresponds to the water transfer (osmosis) from the posolyte to the negolyte during the cycling test that has diluted the negolyte and thus, its redox signature intensity.

[0216] Further analyses were performed using .sup.1H NMR spectroscopy, see FIG. 8.

[0217] The protons signatures are similar. The aromatic protons are situated between 6.6 and 7.2 ppm. The protons attached to the methyl group are located at 2.5 ppm. No additional peaks are observed. This result suggests that the compound B does not undergo any degradation during the cycling test.

[0218] Considering these analyses, the compound B allows to reach high nominal current densities (120 mA/cm.sup.2) with rather low energy densities (2.1 Wh/L). A small and steady capacity loss of ?0.025% per cycle is recorded but is associated to air tightness failure (low electrolyte volumes and small tanks) and/or reversible disproportionation reaction (see 2,6-dihydroanthraquinone behavior in alkaline environment, M. J. Aziz et al., J. Am. Chem. Soc., 2019, 141, 20, 8014-8019).

Compound B-M3CH

Cycling Test

[0219] This cell test was performed using materials described above and with the following parameters: [0220] Active area: 25 cm.sup.2 [0221] Membrane: Aquivion? with 50 microns thickness [0222] Flowrate: 100 mL/min [0223] Negative pole: 40 mL of negolyte comprising 0.1 M of compound B-M3CH, 0.225M KOH, 0.042 M NaOH, 0.3M KCl and 0.05M of NaCl [0224] Positive pole: 44 mL of posolyte comprising 0.15 M potassium ferrocyanide, 0.05 M sodium ferrocyanide, 0.075M KOH and 0.025 M of NaOH [0225] Averaged pH: 13.0 [0226] Cycling test: power static mode [0227] Voltage limits: under power static mode, the lower voltage limit was 0.7 V. The upper voltage limit was fixed at 1.25 V which and was followed by a CV step at the same voltage until 0.25 W limit is reached.

[0228] FIG. 9 presents 100 typical cycles of this battery test among more than 1000 cycles.

[0229] FIG. 9 is a graph representing (squares) the capacity retention, (diamonds) coulombic and (triangles) energy efficiencies of a cell test involving a negolyte with compound B-M3CH. 100 cycles are represented, the power densities in discharge and charge were fixed at 40 mW/cm.sup.2.

[0230] The experimental capacity achieved was 70% at the beginning of the cycling test whereas a steady capacity loss of 0.0099% per cycle was observed. High average coulombic efficiencies were recorded during these 500 cycles (?99.41%) with high (?81.17%) average energy efficiencies.

[0231] Air sealing problematics were encountered during this battery test (due to low electrolyte volumes in small tanks that are difficult to perfectly seal), which is partly responsible for the observed capacity loss (we excluded cross-over, electrolysis or irreversible chemical degradation problematics).

[0232] Furthermore, at 50% SOC, the ASR was tracked during the cycling test using standard procedure. The ASR was quite stable, at 2.0 ?.Math.cm.sup.2, which was consistent with an electrochemical cell comprising a Nafion-type membrane with a thickness of 50 ?m which diffuses a mixture of potassium and sodium cations (alkaline medium).

Post-Cycling Analyses

[0233] A preliminary electrochemical test was performed on the negolyte before and after the cycling test (FIG. 10) which lasted for more than 1000 cycles, or 25 days of constant cycling.

[0234] The typical electrochemical signature of compound B-M3CH was registered before the cycling test. The redox signal was centered at ?0.75 V vs Ag/AgCl (or ?0.52 V vs NHE) with a max current of roughly ?2.3 mA during reduction step and a max current of roughly 2.0 mA during reoxidation step. After cycling, the redox signature corresponding to the compound B-M3CH has increased in current intensity, about 30%, which corresponds to the water transfer (osmosis) from the negolyte to the posolyte during the cycling test that has concentrated the negolyte and thus, increased the redox signature intensity of B-M3CH compound.

[0235] Further analyses were performed using .sup.1H NMR spectroscopy, see FIG. 11.

[0236] The protons signatures are similar. The aromatic protons are situated between 6.2 and 7.3 ppm. The protons attached to the (CH2).sub.3 alkyl chain are located at 1.8-1.9, 2.2-2.3 and at 3.5-3.7 ppm. No additional peaks are observed. This result suggests that the compound B-M3CH does not undergo any degradation during the cycling test.

[0237] Considering these analyses, the compound B-M3CH allows to reach good nominal current densities (40 mA/cm.sup.2) with rather low energy densities (2.0 Wh/L). A steady capacity loss of ?0.0099% per cycle is recorded but is associated to air tightness failure (low electrolyte volumes and small tanks). This result shows that the chemical modification of compound B into compound B-M3CH has increased its stability in the application. This effect can be explained by the increase of redox potential of the anthraquinone though the alkylation of one phenolic group that may stabilize of its reduced state (see 2,6-dihydroanthraquinone behavior in alkaline environment, M. J. Aziz et al., J. Am. Chem. Soc., 2019, 141, 20, 8014-8019).