REDOX FLOW BATTERY AND AQUEOUS-BASED SOLUTION

Abstract

The present invention relates to a redox flow battery comprising: a positive compartment containing a positive electrode in contact with a first aqueous-based electrolyte solution comprising a positive electrolyte dissolved in a first aqueous-based solvent; and a negative compartment containing a negative electrode in contact with a second aqueous-based electrolyte solution comprising a negative electrolyte being an organic redox-active molecule dissolved in a second aqueous-based solvent wherein at least one of the first and second aqueous-based electrolytes is based on an ammonium-based salt, wherein the organic redox-active molecule is a naphthalene diimide, abbreviated NDI, or a modified NDI.

Claims

1. A redox flow battery comprising: a positive compartment comprising a positive electrode in contact with a first aqueous-based electrolyte solution comprising a positive electrolyte dissolved in a first aqueous-based solvent; a negative compartment comprising a negative electrode in contact with a second aqueous-based electrolyte solution comprising a negative electrolyte being an organic redox-active molecule dissolved in a second aqueous-based solvent; electrical conductive means for establishing electrical conduction between said positive electrode and said negative electrode, and an external load for directing electrical energy into or out of the redox flow battery; a separator component that separates the first aqueous-based electrolyte solution in the positive compartment from the second aqueous-based electrolyte solution in the negative compartment and substantially prevents the positive electrolyte in the positive compartment and the negative electrolyte in the negative compartment from intermingling with each other, while permitting the passage of non-redox-active species between the electrolyte solutions in the positive and negative compartments; and means capable of establishing flow of the electrolyte solutions past said positive and negative electrodes, respectively, wherein at least one of the first and second aqueous-based electrolytes is based on an ammonium-based salt, and wherein the organic redox-active molecule is a naphthalene diimide, abbreviated NDI, or a modified NDI.

2. A redox flow battery according to claim 1, wherein the ammonium-based salt comprises at least one of the following: ammonium chloride, ammonium phosphate.

3. A redox flow battery according to claim 1, wherein the other one of the first and second aqueous-based electrolytes is based on an ammonium salt, preferably the same ammonium salt.

4. A redox flow battery according to claim 1, wherein the organic redox-active molecule has a solubility at room temperature of at least 0.4 M in the second aqueous-based electrolyte solution.

5. A redox flow battery according to claim 1, wherein the second aqueous-based electrolyte solution comprises at least two different organic redox-active molecules dissolved in the second aqueous-based solvent, a first organic redox-active molecule being NDI, and a second organic redox-active molecule being modified NDI.

6. A redox flow battery according to claim 1, wherein the modified NDI is a substituted NDI, e.g. a core-aminated NDI.

7. A redox flow battery according to claim 6, wherein the modified NDI comprises an amino group.

8. A redox flow battery according to claim 1, wherein the modified NDI has a structure according to formula I: ##STR00008## wherein each, of R.sup.1, R.sup.2, R.sup.3, R.sup.4, R.sup.5 and R.sup.6, is independently selected from: hydrogen atom; cyano group (—CN); amino group —NR8R9R10, wherein R10 is present when the amino group is quaternized, and wherein R8, R9 and R10, if R10 present, are independently selected from hydrogen atom and hydrocarbyl group R11 having one to six carbon atoms, or, R8 and R9 forming together with the nitrogen a hetero ring having four to six carbon atoms; sulfonate (—S(O).sub.2OH or —S(O).sub.2O.sup.−); halogen atoms (e.g. fluorine, chlorine, bromine, and/or iodine); and group R7, being a hydrocarbyl group having one to twenty carbon atoms, and having one or more heteroatoms selected from oxygen, nitrogen, sulfur, and halogen atoms (e.g. fluorine, chlorine, bromine, and/or iodine), or group R7, being a hydrocarbyl group having one to twenty carbon atoms, and being substituted by one, two, or three substituents selected from: amino group —NR.sup.8R.sup.9R.sup.10, wherein R.sup.10 is present when the amino group is quaternized, and wherein R.sup.8, R.sup.9 and R.sup.10, if R.sup.10 present, are independently selected from hydrogen atom and hydrocarbyl group R.sup.11 having one to six carbon atoms, or R.sup.8 and R.sup.9 forming together with the nitrogen, or together with the nitrogen and a further nitrogen in either of R.sup.8 or R.sup.9, a hetero ring having four to six carbon atoms, and sulfonate (—S(O).sub.2OH or —S(O).sub.2-O.sup.−) or (—COOH or or COO—).

9. A redox flow battery according to claim 8, wherein R1 and R4 comprise tertiary or quaternary amines, and/or sulfonate groups.

10. A redox flow battery according to claim 8, wherein at least one of R2, R3, R5 and R6 is a group or molecule comprising more than a hydrogen atom, or is different to a hydrogen atom, or wherein at least one of R2, R3, R5 and R6 is an amino or cyano group.

11. A redox flow battery according to claim 1, wherein the battery is configured such that the NDI, or modified NDI, is reduced with two electrons in the negative compartment, creating an NDI dianion or hydroNDI.

12. A redox flow battery according to claim 11, wherein the NDI dianon or reduced NDI is an original reduced NDI having a first structure, and wherein the battery is configured such that the original reduced NDI is restructured into a restructured reduced NDI having a second structure different from said first structure, the restructured reduced NDI having a different reduction potential compared to the original reduced NDI.

13. A redox flow battery according to claim 12, wherein the difference in reduction potential between the original reduced NDI and the restructured reduced NDI determines the voltage of the battery.

14. A redox flow battery according to claim 1, wherein the pH at the negative compatment is lower compared to the positive compartment with at least a value of pH 2.

15. A redox battery according to claim 1, wherein the pH is adjusted during cycling due to the proton-coupled electron transfer of NDI.

16. A redox flow battery according to claim 1, wherein the positive electrolyte is the same as the negative electrolyte, forming a symmetrical redox flow battery.

17. A redox flow battery according to claim 16, wherein the battery is configured such that the NDI, or modified NDI, is reduced with two electrons in the negative compartment, creating an NDI dianion or hydroNDI, wherein the NDI dianon or reduced NDI is an original reduced NDI having a first structure, and wherein the battery is configured such that the original reduced NDI is restructured into a restructured reduced NDI having a second structure different from said first structure, the restructured reduced NDI having a different reduction potential compared to the original reduced NDI, and wherein charging of the battery results in the following reactions: in the negative compartment: NDI+2e.sup.−.fwdarw.NDI.sub.2− and NDI.sub.2−.fwdarw.NDI*.sup.2− or NDI+2e.sup.−+2H.sup.+.fwdarw.NDIH.sub.2 and NDIH.sub.2.fwdarw.NDIH.sub.2*, wherein NDI*.sup.2− and NDIH.sub.2* are forms of the restructured reduced NDI; in the positive compartment: NDI*2−.fwdarw.NDI*+2e−and NDI*.fwdarw.NDI or NDIH2*.fwdarw.NDI*+2e−+2H+ and NDI*.fwdarw.NDI, wherein NDI* is the oxidised condition of the restructured reduced NDI.

18. An aqueous solution comprising NDI or modified NDI having a structure according to formula II: ##STR00009## wherein R.sup.22 and R.sup.25 are both hydrogen atom, and each, of R.sup.21, R.sup.23, R.sup.24 and R.sup.26, is independently selected from: hydrogen atom (H); cyano group (CN); amino group —NR.sup.28R.sup.29R.sup.30, wherein R.sup.30 is present when the amino group is quaternized, and wherein R.sup.28, R.sup.29 and R.sup.30, if R.sup.30 present, are independently selected from hydrogen atom and hydrocarbyl group R.sup.31 having one to four carbon atoms, or R.sup.28 and R.sup.29 forming together with the nitrogen a hetero ring having four to six carbon atoms; sulfonate (—S(O).sub.2OH or —S(O).sub.2O.sup.−) (—COOH or or COO—); halogen atoms (e.g. fluorine and/or bromine); and group R.sup.27, being a hydrocarbyl group having one to six carbon atoms, and having one or more heteroatoms selected from nitrogen, sulfur, and halogen atoms (e.g. fluorine and/or bromine), or group R.sup.27, being a hydrocarbyl group having one to six carbon atoms, and being substituted by one or two substituents selected from: cyano group (—CN); amino group —NR.sup.28R.sup.29R.sup.30, wherein R.sup.30 is present when the amino group is quaternized, and wherein R.sup.28, R.sup.29 and R.sup.30, if R.sup.30 present, are independently selected from hydrogen atom and hydrocarbyl group R.sup.31 having one to four carbon atoms, or R.sup.28 and R.sup.29 forming together with the nitrogen, or together with the nitrogen and a further nitrogen in either of R.sup.28 or R.sup.29, a hetero ring having four to six carbon atoms; and sulfonate (—S(O).sub.2OH or —S(O).sub.2O.sup.−) or (—COOH or or COO—), wherein the aqueous solution further comprises an aqueous-based electrolyte based on an ammonium-based salt.

19. An aqueous solution according to claim 18, wherein each of R21 and R24, is independently selected from: group R.sup.27, being a hydrocarbyl group having one to six, e.g. one to four, carbon atoms, and having one or more, e.g. one or two, heteroatoms selected from nitrogen, sulfur, and halogen atoms (e.g. fluorine and/or bromine), or group R.sup.27, being a hydrocarbyl group having one to six, e.g. one to four, carbon atoms, and being substituted by one or two, e.g. one, substituents selected from: cyano group (—CN); amino group —NR.sup.28R.sup.29R.sup.30, wherein R.sup.30 is present when the amino group is quaternized, and wherein R.sup.28, R.sup.29 and R.sup.30, if R.sup.30 present, are independently selected from hydrogen atom and hydrocarbyl group R.sup.31 having one to four carbon atoms, or R.sup.28 and R.sup.29 forming together with the nitrogen a hetero ring having four to six carbon atoms; and sulfonate (—S(O).sub.2OH or —S(O).sub.2O.sup.−), and wherein R.sup.23 and R.sup.26, is independently selected from: cyano group (CN); amino group —NR.sup.28R.sup.29R.sup.30, wherein R.sup.30 s present when the amino group is quaternized, and wherein R.sup.28, R.sup.29 and R.sup.30, if R.sup.30 present, are independently selected from hydrogen atom and hydrocarbyl group R.sup.31 having one to four carbon atoms, or R.sup.28 and R.sup.29 forming together with the nitrogen a hetero ring having four to six carbon atoms; sulfonate (—S(O).sub.2OH or —S(O).sub.2O.sup.−); halogen atoms (e.g. fluorine and/or bromine); and group R.sup.27, being a hydrocarbyl group having one to six carbon atoms, and having one or more heteroatoms selected from nitrogen, sulfur, and halogen atoms (e.g. fluorine and/or bromine), or group R.sup.27, being a hydrocarbyl group having one to six carbon atoms, and being substituted by one or two substituents selected from: cyano group (—CN); amino group —NR.sup.28R.sup.29R.sup.30, wherein R.sup.30 is present when the amino group is quaternized, and wherein R.sup.28, R.sup.29 and R.sup.30, if R.sup.30 present, are independently selected from hydrogen atom and hydrocarbyl group R.sup.31 having one to four carbon atoms, or R.sup.28 and R.sup.29 forming together with the nitrogen a hetero ring having four to six carbon atoms; and sulfonate (—S(o).sub.2OH or —S(O).sub.2O.sup.−).

Description

BRIEF DESCRIPTION OF DRAWINGS

[0207] The invention will be described in greater detail in the following, with reference to the embodiments that are shownin the attached drawings, in which:

[0208] FIG. 1 shows a titration curve of 50 ml 15.6 mM tertiary NDI in water by addition of 1 M sodium hydroxide.

[0209] FIGS. 2a and 2b show cyclic voltammograms (CVs) for 1 mM NDI in a) neutral phosphate buffer and b) 1 M H.sub.2SO.sub.4.

[0210] FIG. 3 shows Potential-pH diagram of NDI. At pH below 3, the second reduction follows a one electron-1.5 proton slope (in total 2 electrons and three protons), and at higher pHs, it is independent of pH, as the first reduction couple is over the whole range.

[0211] FIG. 4 shows rotating disk electrode (RDE) analysis for 1 mM NDI in 1 M H.sub.2SO.sub.4.

[0212] FIG. 5 shows RDE for NDI in pH 7 phosphate buffer solution at concentrations of a), b) 1 mM and c), d) 50 mM.

[0213] FIG. 6 shows bulk electrolytic cycling of 25 mM NDI in pH 6.4 phosphate buffer with a phosphate total concentration of 0.5 M.

[0214] FIG. 7 shows cyclic voltammograms comparing the redox activity of the solution before and after 12 cycles in the bulk electrolysis cell.

[0215] FIG. 8 shows Cyclic voltammograms of “the compound obtained in Example 3” measured versus Ag/AgCl: in 1 M H.sub.2SO.sub.4 at different scan rates.

[0216] FIG. 9 shows Cyclic voltammograms of “the compound obtained in Example 3” measured versus Ag/AgCl: in 0.5 M sodium phosphate buffer pH 7 at different scan rates.

[0217] FIG. 10 shows Cyclic voltammograms of “the compound obtained in Example 4b” measured versus Ag/AgCl: in 1 M H.sub.2SO.sub.4 at different scan rates.

[0218] FIG. 11 shows Cyclic voltammograms of “the compound obtained in Example 4b” measured versus Ag/AgCl: in 0.5 M sodium phosphate buffer pH 7 at different scan rates.

[0219] FIG. 12 shows a schematic view of an aqueous quinone-based redox flow battery.

[0220] FIG. 13 shows a cyclic voltammograms of 15 mM partially reduced NDI in 1 M sulfuric acid solution. Sweep rate=100 mV/s.

[0221] FIG. 14 shows bulk electrolysis cycling of 15 mM NDI in 1 M sulfuric acid solution.

[0222] FIG. 15 shows the capacity utilization for four different redox flow batteries based on different NDIs in combination with different aqueous based electrolytes.

DETAILED DESCRIPTION

[0223] The embodiments of the invention with further developments described in the following are to be regarded only as examples and are in no way intended to limit the scope of the protection provided by the patent claims.

EXPERIMENTAL SECTION

EXAMPLES

Example 1

Naphthalene Diimide (NDI)

[0224] NDI was synthesized in 30 g scale according to Sissi, C. et al, Bioorg. Med. Chem. 2007, 15, (1), 555-62, in a facile and green manner with high yields and water as the only byproduct, and the quaternization, if necessary, simply by bubbling chloromethane through a chloroform solution of the tertiary NDI, whereupon the pure product is precipitated, see Scheme.

##STR00003##

[0225] The quaternization is only necessary if the molecule is to be examined at higher pH values, since the tertiary amine sidechains are protonated for pH values less than about 8 (pKa≈8.1) as shown by the titration curve in FIG. 1. In FIG. 1 it can be seen that the molecule deprotonates above pH 8 and precipitates out of solution. For actual flow battery applications, i.e. redox flow battery according to the present invention, the last reaction step is thus possible to omit.

Electrochmistry

Cyclic Voltammetry (CV)

[0226] The electrochemistry of NDI was examined by CV at a range of different pH values, to gain understanding of the electron transfer reaction mechanism. FIG. 2a and FIG. 2b show cyclic voltammograms (CVs) for NDI at pH 7 and 0, respectively. An initial observation shows that more than one electron is involved, due to the broadness of the peak at pH 0, and due to the occurrence of two separate redox couples at pH 7. The broadness of the first reduction at pH 7, is assumed to be due to the possibility for the first electron to add at different positions of NDI.

[0227] A plot of the reduction potentials for NDI at different pHs is seen in FIG. 3. The stars show that the first reduction potential is independent of pH and does not involve any protons. The second reduction, the triangles in the figure, has a slope of 91 mV/pH at pH below 3, indicating a one electron-1.5 proton relationship, according to the Nernst equation.

[0228] It should be remarked that the reduction potential of NDI is outstandingly low, and if used in a redox flow battery, it would possibly yield an operating voltage among the highest of the previously reported structures in aqueous solutions. Indeed, more negative potentials are not beneficial, due to the parasitic hydrogen evolution reactions eventually taking precedence.

[0229] Further, rotating disk electrode (RDE) voltammetry was coupled with diffusion NMR to assess the accessible concentration of NDI in a neutral buffered solution, and tested whether it corresponded to dimerization. The diffusion coefficients from diffusion NMR were found to be 2.57 and 2.34×10.sup.−6 cm.sup.2 s.sup.−1 for the acid and neutral solutions respectively. From the slope in FIG. 4, the accessible concentration was calculated, through the Levich equation, and showed that the reduction of 1 mM NDI at pH 0 only involves two electrons. Addition of three protons to each NDI molecule at pH values as high as 3 is unlikely, since the third proton would be at a weakly basic carbonyl. This is further supported by the lack of buffering at lower pHs in the titration curve in FIG. 1.

[0230] Based on this relationship and reasoning, the reduction scheme in Scheme 2 is proposed for acidic solutions as follows.

##STR00004##

Dimerization

[0231] As work on 9,10-Anthraquinone-2,7-disulfonic acid (AQDS) has shown the impact of dimerization on its electrochemical properties, it was prudent to examine whether the same behavior could be seen for NDI. Since NDI has a large aromatic core, and it is known to self-associate quite readily, it was imagined that the effect could be quite strong, despite the flexible positively charged sidechains.

[0232] FIG. 5 shows the results from RDE measurements of NDI in neutral buffered solution. Two well-defined diffusion-limited current plateaus were observed, with the first reduction coming at a less negative potential, and the second reduction at a more negative potential, for 50 mM compared to 1 mM. The large distance between reduction potentials could enable the use of NDI in a symmetric flow battery, where NDI is utilized as both cathodic and anodic electrolytes simultaneously, but has not been further examined up to this point. The diffusion-limited currents were fitted to the Levich equation along with the diffusion coefficient acquired from diffusion NMR, showing a linear relationship. The slope of the fit showed, through the Levich equation, that the full capacity is not accessed on the voltammetry timescale, possibly due to dimerization. However, as seen in the next section, the full capacity can be accessed through bulk electrolysis, and it is thought that the dimer dissociates too fast to impede the redox reaction in the flow battery application.

Bulk Electrolysis

[0233] A solution of 25 mM NDI in pH 6.4 sodium phosphate buffer was cycled in a bulk electrolysis cell, where all of the material is reduced, as opposed to CV and RDE, where only a small fraction of the material is probed. A galvanostatic BE setting was used, where a constant current of 40 mA was applied. The electrolysis was intermittently paused, allowing for CVs to be collected to be able to quantify the amount of NDI that has been reduced. The peak heights from the CVs were correlated to the passed current, and it was seen that the complete capacity of NDI could be accessed using bulk electrolysis. See here FIG. 6 which shows bulk electrolytic cycling of 25 mM NDI in pH 6.4 phosphate buffer with a phosphate total concentration of 0.5 M. A possible explanation to why NDI can be quantitatively reduced with BE, and AQDS cannot, is the formation of the AQDS quinhydrone upon reduction, where the reduced species associates with the pristine molecule, to form a new self-association complex, which seems to, similarly to the AQDS dimer, impede the reduction. For NDI, the reduction of the molecule lowers the monomer concentration in the solution, and the dimerization equilibrium is shifted in the direction of dimer separation. Two plateaus were seen at the electrolysis, corresponding to the two separate redox couples, illustrated by the overlaid CV in FIG. 6.

[0234] After the initial sweep, continuous cycling at a current of 15 mA was performed for twelve cycles. A CV was collected after the cycling had finished, roughly a week after starting. The CV was compared to the initial, untouched solution, and it was seen that they overlapped perfectly, indicating no or little material degradation over the course of the experiment, showing the exceptional stability of NDI. See here FIG. 7 which shows cyclic voltammograms comparing the redox activity of the solution before and after 12 cycles in the bulk electrolysis cell. Thus, bulk electrolysis shows (see FIG. 6 and FIG. 7) that NDI does not degrade at all, after being cycled 12 times over a week's time. Apart from that, the NDI molecule is synthesized via a one-step reaction, at room temperature in water with an almost 100% yield. It is water soluble up to 0.7 M, and has a lower reduction potential than any of the other examined molecules, resulting in a high battery voltage in the application.

[0235] The cost and performance of the aqueous redox flow battery depends very strongly on the redox-active material employed. The starting material for the synthesis of NDI, 1,4,5,8-naphthalene tetracarboxylic acid dianhyride is easily synthezied from naphthalene—a very cheap and abundant compound. The combination of having a very low reduction potential, high aqueous solubility, high chemical stability, low membrane permeability, environmental benignity, possibility for use at neutral pH, and potential for manufacture at very low costs, is a set of criteria that is not easily fulfilled. NDI achieves a fulfillment of these criteria to a far higher extent than competing technologies. Thus, a flow battery employing NDI has a superior cost-performance relationship than competing technologies, without compromising on environmental impact or operational safety.

Example 2

2,6-dibromonaphtalene-1,4,5,8-tetracarboxylic dianhydride

[0236] A 250 ml round-bottom flask charged with 1,4,5,8-naphthalene tetracarboxylic acid dianhydride (NDA) (4.00 g, 14.92 mmol) and sulfuric acid (98%, 100 ml) was subjected to stirring at 50° C. for an hour to enhance dissolution. To the resulting solution was added by portions dibromoisocyanuric acid (DBI) (6.46 g, 22.51 mmol) and stirred at ambient temperature for 30 minutes. The reaction temperature was then adjusted and maintained at 120° C. for 45 hours.

[0237] Upon heating, reddish-brown fumes evolved which were trapped in a solution of sodium thiosulphate. The reaction mixture was cooled to room temperature, remaining reddish fumes were removed with a gentle stream of nitrogen gas and then poured into crushed ice. The precipitate formed was filtered, washed with copious amount of distilled water and then methanol. Br,Br-NDA was obtained as a yellowish-green solid and vacuum-dried at 40° C. overnight to afford approximately 78% yield. This was used without any further purification.

[0238] The nuclear magnetic resonance (NMR) spectroscopy for 2,6-dibromonaphtalene-1,4,5,8-tetracarboxylic dianhydride is: .sup.1H NMR (400 MHz, DMSO-d6) δ 8.77 (s, 2H); .sup.13C NMR (101 MHz, DMSO-d6) δ 158.39, 156.89, 137.97, 129.87, 127.85, 124.72, 123.87.

Example 3

4,9-dibromo-2,7-bis(3-(dimethylamino)propyl)benzo[Imn][3,8]phenanthroline-1,3,6,8(2H,7H)-tetraone

[0239] 2,6-dibromonaphtalene-1,4,5,8-tetracarboxylic dianhydride from Example 2 (5.01 g, 11.76 mmol) and 50 ml of glacial acetic acid were charged into a two-necked round-bottom flask under stirring. To the system was added 3-(dimethylamino)-1-propylamine (3.61 g, 4.45 ml, 35.33 mmol, 3 eq). The mixture was stirred at 120° C., and after 30 minutes of reaction, cooled to room temperature, quenched in ice, neutralized with sodium carbonate and extracted three times with chloroform. The organic layer was dried under vacuum to give an orange crude, which was purified by column chromatography (CHCl3:MeOH, 9:1) to obtain approximately 20% yield. The silica stationary phase was pretreated with triethylamine.

[0240] The nuclear magnetic resonance (NMR) spectroscopy for 4,9-dibromo-2,7-bis(3-(dimethylamino)propyl)benzo[Imn][3,8]phenanthroline-1,3,6,8(2H,7H)-tetraone is: .sup.1H NMR (400 MHz, Chloroform-d) δ 8.98 (s, 2H), 4.31-4.22 (m, 4H), 2.45 (t, J=7.0 Hz, 4H), 2.24 (s, 12H), 1.99-1.85 (m, 4H); .sup.13C NMR (101 MHz, Chloroform-d) δ 160.79, 160.75, 138.99, 128.28, 127.70, 125.32, 124.10, 57.06, 45.23, 39.90, 25.54.

Example 4a and 4b

Benzo[Imn][3,8]phenanthroline-2,7-dipropanaminium, 1,3,6,8-tetrahydro-N,N,N,N′,N′,N′-hexamethyl-1,3,6,8-tetraoxo-, dichloride (Example 4a) and 3,3′-(4,9-dibromo-1,3,6,8-tetraoxo-1,3,6,8-tetrahydrobenzo[Imn][3,8]phenanthroline-2,7-diyl)bis(N,N,N-trimethylpropan-1-aminium) chloride (Example 4b)

[0241] A 100 ml round-bottom flask was charged at each reaction with 50 mg of either NDI or Br,Br-NDI for the synthesis of compound of Example 4a and compound of Example 4b, respectively. To the system was added 10-15 ml of chloroform and subjected to heating in an oil bath at 50° C. while stirring. Chloromethane gas was bubbled through the content of the flask three times for at most a minute each time. Precipitates were observed to form after an hour of stirring. The system was left to stir for 12 hours. The reaction was cooled to room temperature, filtered and dried under vacuum to afford a quantitative yield.

[0242] The compound of Example 4a was obtained as an off-white solid, and the nuclear magnetic resonance (NMR) spectroscopy for Benzo[Imn][3,8]phenanthroline-2,7-dipropanaminium, 1,3,6,8-tetrahydro-N, N, N,N′,N′, N′-hexamethyl-1,3,6,8-tetraoxo-, dichloride is:

[0243] .sup.1H NMR (400 MHz, Deuterium Oxide) δ 8.41 (s, 4H), 4.09 (t, J=7.0 Hz, 4H), 3.44-3.35 (m, 4H), 3.02 (s, 18H), 2.20-2.07 (m, 4H); .sup.13C NMR (101 MHz, Deuterium Oxide) δ 163.88, 130.96, 125.86, 125.81, 64.00, 52.88, 37.63, 21.34.

[0244] The compound of Example 4b was obtained as an orange-red solid and the nuclear magnetic resonance (NMR) spectroscopy for 3,3′-(4,9-dibromo-1,3,6,8-tetraoxo-1,3,6,8-tetrahydrobenzo[Imn][3,8]phenanthroline-2,7-diyl)bis(N,N,N-trimethylpropan-1-aminium) chloride is:

[0245] .sup.1H NMR (400 MHz, Deuterium Oxide) δ 8.56 (s, 2H), 4.07 (t, J=7.0 Hz, 4H), 3.44-3.36 (m, 4H), 3.02 (s, 18H), 2.15-2.06 (m, 4H). .sup.13C NMR (101 MHz, Deuterium Oxide) δ 161.21, 138.48, 128.22, 126.52, 124.31, 123.31, 63.85, 63.82, 63.79, 55.15, 52.95, 52.91, 52.87, 42.74, 38.33, 21.14.

Solubility Testing

[0246] A saturated solution of about 500 μl each of the targeted molecules, i.e the compounds obtained in Examples 3, 4a and 4b, in milli-Q water containing undissolved particles were prepared at ambient temperature. The solutions were centrifuged at 12,500 rpm for 30 minutes. 100 μl of the supernatant were pipetted in triplicates from each of the solution into separate empty vials of known weights. The content of these vails were left to dry in a vacuum oven overnight. The final weights of the vials were determined and the amount of each of the compounds calculated from the weight difference. Solubility in mmol/L for each compound was finally determined. The results obtained are presented in Table 1, where “Compound 1” is the compound obtained in Example 3, “Compound 2” is the compound obtained in Example 4a and “Compound 3” is the compound obtained in Example 4b. It should be noted that the solubility may be increased by adding a buffer or acid. For example, by adding a buffer or acid to compound 1, the solubility is increased above 0.4 M.

TABLE-US-00001 TABLE 1 Solubility (mmol/L) Compound Test 1 Test 2 Test 3 Average 1 35.97 34.06 35.59 35.21 2 653.42 689.14 691.00 677.85 3 142.39 151.02 154.18 149.20

[0247] Solubility of the compound obtained in Example 3 which is a base NDI is lower than those of the compounds obtained in Examples 4a and 4b which are quaternary ammonium NDI salts. This is due to the differences in intermolecular interactions of the individual compounds. The compound obtained in Example 3 contains soluble amines both at the imide positions and the naphthalene core eventually affects the solubility at the molecular level. Adversely, a long range intermolecular π-π stacking effect of the molecules in solution could not be hindered, therefore limiting their solubility. Since the compounds obtained in Examples 4a and 4b exist as salts, they could dissolve more easily in water due to electrostatic interactions of their resulting ions. This could limit the π-π stacking of the NDI to an extent of permitting higher solubilities than compound 1. The difference in the solubilities of the compounds obtained in Examples 4a and 4b is resulting from the insoluble bromine substituent on the naphthalene core. If the structure obtained in Example 3 was to be added to a slightly acidic solution, the amines would become protonated and the solubility would increase significantly. The structures of the compounds obtained in Examples 3, 4a and 4b are shown in Scheme 3.

##STR00005##

Electrochemical Measurements

[0248] All CV measurements were carried out in a custom built 10 ml three-electrode electrochemical cell. The lower part of the cell holds the electrolyte of interest. In the upper part are ports to hold the working electrode, reference electrode and auxiliary electrode as well as inlet and outlet for N.sub.2 gas. 1 mM solutions of compounds, i.e the compounds obtained in Examples 3, 4a and 4b, each in either 1 M sulfuric acid or 0.5 M sodium phosphate buffer of pH 7 electrolytes. Cyclic voltammograms were recorded at scan rates of 10 mVs.sup.−1, 20 mVs.sup.−1, 50 mVs.sup.−1, 100 mVs.sup.−1 and 250 mVs.sup.−1 at room temperature. Prior to measurements, the electrolyte was de-aerated by continuously purging with N.sub.2 gas for 10-15 minutes and maintaining a N.sub.2 flow-blanket throughout the experiment to minimize any environmental contamination. Also, 80% of the ohmic drop was compensated for during the experiment using a positive feedback-loop.

CV Analysis

The Compound Obtained in Example 3

[0249] FIG. 8 and FIG. 9 provides CV reports of “the compound obtained in Example 3” in acidic and neutral electrolytes at different scan rates showing single reduction and reversible oxidation peaks. Corresponding numerical data are as summarized in Tables 2 and 3. The redox potential, E.sup.0′, obtained was −0.13 V vs. Ag/AgCl in the acidic electrolyte and −0.43 V vs. Ag/AgCl in the neutral electrolyte. At scan rates of 10 mVs.sup.−1, 20 mVs.sup.−1 and 50 mVs.sup.−1, the compound revealed comparatively low peak-to-peak potential separations, ΔE.sub.p, in the neutral electrolyte than in the acidic electrolyte. Nonetheless, variation of ΔE.sub.p with increasing scan rate was more pronounce in the neutral electrolyte than in the acidic electrolyte as shown in the respective tables. The results are suggestive that “the compound obtained in Example 3” is generally electrochemically reversible although the AEp are observed to drift with increasing scan rates. Electrochemical reversibility of this compound was further established by calculating peak current ratios, i.sub.pa/i.sub.pc, at the various scan rates as shown in the Tables 2 and 3. For an ideal reversible process, this parameter should be unity. The values as shown in the tables indicate that “the compound obtained in Example 3” exhibited good reversibility in the acidic electrolyte than in the neutral electrolyte. The varying CV data in both electrolytes predict that the redox chemistry of this compound can be altered by the choice of electrolyte.

TABLE-US-00002 TABLE 2 CV data of “the compound obtained in Example 3” in 1M sulfuric acid at different scan rates Scan Rate (mVs.sup.−1) E.sup.0′ vs. Ag/AgCl (V) ΔE.sub.p (mV) [00001]$\frac{i_a}{i_c}$ 10 −0.13 63.67 ± 10.03 0.91 ± 0.05 20 −0.13 63.31 ± 10.97 0.92 ± 0.01 50 −0.13 64.97 ± 10.46 0.89 ± 0.02 100 −0.13 66.31 ± 9.29 0.87 ± 0.00 250 −0.13 71.25 ± 11.85 0.82 ± 0.01

TABLE-US-00003 TABLE 3 CV data of “the compound obtained in Example 3” in 0.5M sodium phosphate buffer of pH 7 at different scan rates Scan Rate (mVs.sup.−1) E.sup.0′ vs. Ag/AgCl (V) ΔE.sub.p (mV) [00002]$\frac{i_a}{i_c}$ 10 −0.43 49.21 ± 8.49 0.82 ± 0.04 20 −0.43 51.87 ± 9.16 0.80 ± 0.05 50 −0.43 61.48 ±12.09 0.74 ± 0.06 100 −0.43 68.59 ± 14.05 0.69 ± 0.07 250 −0.43 86.49 ± 20.99 0.63 ± 0.07

The Compounds Obtained in Examples 4a and 4b

[0250] The CVs for the compounds obtained in Examples 4a and 4b revealed two pseudo-reversible redox processes in both the acidic and neutral electrolytes, see CVs for the compounds obtained in Example 4b displayed in FIG. 10 and FIG. 11. The shapes of the CV waves do not give a clear separation between the different redox processes thus, transfer from one complete redox process to the other as the potential ramps does not show a good exponential decay of the peak currents which makes evaluation of the individual processes appear complex. There was no evidence of decomposition observed for these compounds in both electrolytes. The results can be suggestive that the compounds exhibit sluggish redox chemistries in the chosen supporting electrolytes.

[0251] The CVs of “the compound obtained in Example 4a” in the acidic electrolyte at 1 mM concentration appeared to give a single but broad reduction and reverse oxidation peaks (CVs not shown). Further investigations were conducted by running CVs for 10 mM and 25 mM concentration of the compound in the acid electrolyte. These then revealed CV waves with two pairs of peaks almost merged together which is characteristic of two separate electron transfer steps with similar redox potentials. When “the compound obtained in Example 4a” was examined in the neutral electrolyte, the resolution of these pairs of peaks become more pronounced (CVs not shown). In the acidic electrolyte, the redox potential, E.sup.0′, was approximated to be −0.32 V vs. Ag/AgCl. In the neutral electrolyte, the two observed redox processes could be approximated to occur at potentials, E.sup.0′, −0.30 V vs. Ag/AgCl and −0.63 V vs. Ag/AgCl.

[0252] The compound obtained in Example 4b gave the most complicated CVs. Two observable pairs of peaks appeared at approximated potentials −0.13 V vs. Ag/AgCl and −0.38 V vs. Ag/AgCl in the acidic electrolyte. In the neutral electrolyte, the redox processes were observed at −0.05 V vs. Ag/AgCl and −0.65 V vs. Ag/AgCl. The broad cathodic and anodic peaks appearing respectively in the first and second redox processes coupled with the high capacitance made it difficult to evaluate peak-to-peak potential separations. Although the voltammograms are complex the systems are in total reversible.

Electrochemical Behaviour of the Target NDIs i.e. “Compound Obtained in Example 3”, “Compound Obtained in Example 4a”, and “Compound Obtained in Example 4b”

[0253] Different electrochemical behaviours were observed for the synthesized NDIs i.e. “compound obtained in Example 3”, “compound obtained in Example 4a”, and “compound obtained in Example 4b” in the acidic and neutral electrolytes. The redox potential of each compound showed dependency on pH. All compounds except for “the compound obtained in Example 3” displayed two redox waves characteristic of two separate electron transfer processes. “The compound obtained in Example 3” showing only one redox wave may be that the two-electron transfer processes occur simultaneously at similar potentials at a relatively fast rate. Effects of the different NDI core-substituents and the different background electrolytes can be seen by the shift of the redox potentials. Although the second redox processes for “compounds obtained in Examples 4a and 4b” appear to occur at the same potential, it is clear that the influences of the core substitutions are pronounced on the first redox process. The potential gap between the first and second redox waves narrows down by pushing the former wave to a more negative redox potential when moving from the electron withdrawing bromo-substituents on “compound obtained in Example 4b” to the electron donating amino-groups on “compound obtained in Example 3”. By comparing the compounds in the acidic electrolytes (red CV waves), “the compound obtained in Example 3” shows a redox potential which is more positive than that of “the compound obtained in Example 4a” but almost at the same potential as the first redox wave of the “compound obtained in Example 4b”. This is because in the acidic electrolyte, the amino substituents at the core of “the compound obtained in Example 3” are protonated causing a switch from their usual electron donating properties to electron withdrawing and eventually leading to a facile shift of the redox wave to a more positive potential than that of “the compound obtained in Example 4a”. In the neutral electrolyte (blue CV waves), the potential gap between the first and second redox processes for “the compounds obtained in Examples 4a and 4b” becomes relatively wider compared to their respective behaviours in the acidic electrolyte. The electron withdrawing bromo-substituents on “the compound obtained in Example 4b” influences the first redox wave to appear at a more positive potential than that of “the compound obtained in Example 4a” as a result of an inductive effect of the bromides.

[0254] FIG. 12 shows a schematic view of an aqueous quinone-based redox flow battery.

[0255] Further, in FIG. 12 the electrolyte solutions in the two tanks contain two different (redox-active molecules) RMs, and the difference in reduction potential between the two molecules determines the open-circuit voltage. The electrolyte solution containing the RM with the more negative redox potential is called the negative electrolyte, and the solution containing the RM with the less negative (or positive) redox potential is called positive electrolyte. A supporting electrolyte serves to make the solution ionically conductive and provide the system with mobile charge carriers.

[0256] During discharge, the electrolyte solutions are pumped through an electrochemical cell consisting of two porous electrodes, which are separated by an ion-selective membrane. The molecules in the positive electrolyte, once they reach the surface of the porous electrode, get oxidized and give off electrons which are conducted through an external circuit and used as electricity. At the same time, the negative electrolyte receives electrons from the porous electrode on the cathodic side, and thus gets reduced. Since electrons have effectively been transported from one side of the cell to the other, a charge imbalance has arisen. To negate this, cation or a proton migrates through the membrane from the anodic (positive) to the cathodic (negative) chamber of the electrochemical cell. Thus, FIG. 12 shows the operation of a quinone-based organic flow battery.

[0257] FIG. 13 shows a cyclic voltammogram of 15 mM partially reduced NDI in a 1 M sulfuric acid solution. The figure shows a large separation between multiple two-electron redox processes, enabling the system for use in a symmetrical flow battery setup. In more details, the symmetrical redox flow battery comprises a negative electrolyte with an NDI molecule (e.g. NDI or modified NDI as previously described) which is reduced with two electrons on the negative side (i.e. in the negative compartment), creating the reduced NDI in the form of the NDI dianion, NDI.sup.2− or hydroNDI, NDIH.sub.2. The symmetrical redox flow battery is configured such that the reduced NDI restructure, e.g. chemically restructure, into a restructured reduced NDI: NDI*.sub.2−, having a different structure different to the reduced NDI (NDI.sup.2−). The restructured reduced NDI (NDI*.sup.2−) has a different reduction potential from the reduced NDI (NDI.sup.2−) and is oxidized at relatively higher potentials than the reduced NDI (NDI.sup.2−). The difference in reduction potential between the reduced NDI (NDI.sup.2−) and the restructured reduced NDI (NDI*.sup.2−) determines the voltage of the battery. The main advantage of the NDI restructuring mechanism is that a twice as high number of electrons is accessible for battery utilization. The following reactions are believe to occur: [0258] on the negative side (in the negative compartment): NDI+2e.sup.−.fwdarw.NDI.sup.2− and NDI.sup.2−.fwdarw.NDI*.sup.2−, wherein NDI*.sup.2− is the restructured reduced NDI (NDI anion) described above; [0259] on the positive side (in the positive compartment): NDI*.sup.2−.fwdarw.NDI*+2e.sup.− and NDI*.fwdarw.NDI, wherein NDI* is the oxidised condition of the restructured redcued NDI.

[0260] FIG. 14 shows a polarization curve of a bulk electrolysis cell containing 15 mM NDI in 1 M sulfuric acid. In more details, the symmetrical redox flow battery comprises a negative electrolyte with an NDI molecule (e.g. NDI or modified NDI as previously described) which is reduced with two electrons on the negative side (i.e. in the negative compartment), creating the hydroNDI species, NDIH.sub.2. The symmetrical redox flow battery is configured such that hydroNDI, NDIH.sub.2, restructures, e.g. chemically restructures, into a restructured hydroNDI, NDIH.sub.2*, having a different structure different to hydroNDI, NDIH.sub.2. The restructured hydroNDI, NDIH.sub.2*, has a different reduction potential from the to hydroNDI, NDIH.sub.2 and is oxidized at relatively higher potentials than hydroNDI, NDIH.sub.2. The difference in reduction potential between the original hydroNDI, NDIH.sub.2 and the restructured hydroNDI, NDIH.sub.2* determines the voltage of the battery. The main advantage of the NDI restructuring mechanism is that a twice as high number of electrons is accessible for battery utilization. The following reactions are believed to occur: [0261] on the negative side (in the negative compartment): NDI+2e.sup.−+2H.sup.+.fwdarw.NDIH.sub.2 and NDIH.sub.2.fwdarw.NDIH.sub.2*, wherein NDIH.sub.2* is the restructured hydroNDI described above; [0262] on the positive side (in the positive compartment): NDIH.sub.2*.fwdarw.NDI*+2e.sup.−+2H.sup.+ and NDI*.fwdarw.NDI, wherein NDI* is the oxidised condition of the restructured hydroNDI, NDIH.sub.2*.

[0263] The NDI may e.g. be dissolved in an acidic electrolyte, for example 1 M sulfuric acid.

[0264] FIG. 15 shows capacity utilization for four different redox flow batteries based on different modified NDIs in combination with different aqueous based electrolytes. The redox flow batteries are for example based on the principle shown in FIG. 12.

[0265] The four different redox flow batteries were assembled according to table 4 below. For all four redox flow batteries, the negative electrolyte was based on NDI (denoted NDI-1) or modified NDI (denoted NDI-2) in a potassium chloride (KCl)-potassium phosphate (KPh) solution or NDI-1 and NDI-2 in an ammonium chloride (AmCl)-ammonium phosphate (AmPh) solution, while the positive electrolyte was based on BTMAP-Fc in a KCl-KPh solution and BTMAP-Fc in an AmCl-AmPh solution. NDI-1 and NDI-2 had the general structure according to formula III with the below listed specification:

##STR00006##

[0266] The BTMAP-Fc had a structure according to formula IV:

##STR00007##

[0267] NDI-1 had the specific formula: 2,7-bis(3-(dimethylamino)propyl)benzo[Imn][3,8]phenanthroline-1,3,6,8(2H,7H)-tetraone and NDI-2 had the specific formula: 4,9-bis(dimethylamino)-2,7-bis(3-(dimethylamino)propyl)benzo[Imn][3,8]phenanthroline-1,3,6,8(2H,7H)-tetraone. The first and second redox flow batteries had NDI-1 and the second and third redox flow batteries had NDI-2. The details of the composition of the first, second, third and fourth redox flow batteries are given in table 4 (the negative aqueous-based electrolyte corresponds to the second aqueous-based electrolyte and the positive aqueous-based electrolyte corresponds to the first aqueous-based electrolyte).

TABLE-US-00004 TABLE 4 Redox flow Negative aqueous- battery no based electrolyte Positive Electrolyte 1 NDI-1 in 1M KCl and BTMAP-Fc in 1M KCl and 0.5M KPh 0.5M KPh 2 NDI-1 in 1M AmCl and BTMAP-Fc in 1M AmCl and 0.5M AmPh 0.5M AmPh 3 NDI 2 in 1M KCl and BTMAP-Fc in 1M KCl and 0.5M KPh 0.5M KPh 4 NDI 2 in 1M AmCl and BTMAP-Fc in 1M AmCl and 0.5M AmPh 0.5M AmPh

[0268] For each one of the four redox flow batteries, 10 ml of the negative electrolyte, and 20 ml of the positive electrolyte was used, to balance the capacities, and the concentration of the redox active materials (BTMAP-Fc, NDI-1 and NDI-2) was 50 mM for all the four redox flow batteries. Each one of the redox flow batteries were cycled (100 cycles) with a current density of 10 mA cm−2, and each cycle took between 1.5 and 2 hours.

[0269] The capacity utilizations of the four redox flow batteries shown in FIG. 15 are calculated based on the following relationship:

[00003]$\mathrm{Capacity}\mathrm{Utilization}=\frac{\mathrm{Discharge}\mathrm{Capacity}}{\mathrm{Theoretical}\mathrm{Capacity}}*100$

[0270] As seen in FIG. 15, a clear trend of diverging capacity utilization factor is seen when comparing the use of ammonium based negative electrolyte (ammonium chloride (AmCl)-ammonium phosphate (AmPh)), i.e. redox flow batteries number 2 and 4 (denoted “Battery 2” and “Battery 4” in FIG. 15), compared to the potassium based negative electrolyte (i.e. potassium chloride (KCl)-potassium phosphate (KPh)), i.e. redox flow batteries number 1 and 3 (denoted “Battery 1” and “Battery 3” in FIG. 15). As a note, the capacity utilization for redox flow battery number 3 fluctuated during the first 60 cycles as a result of a flow obstruction in the electrolyte reservoir. When corrected, the above trend is clearly shown for the following 40 cycles.

[0271] In summary, a significant capacity loss is observed for redox flow batteries number 1 and 3 over 100 cycles, while redox flow batteries number 2 and 4 instead show a slow capacity increase. Thus, the impact of the ammonium cation on the cycling stability is apparent for both NDI-1 and NDI-2.