The oxidation of organic additives in the positive vanadium electrolyte and its effect on the performance of vanadium redox flow battery

Journal of Power Sources 334 (2016) 94e103

Contents lists available at ScienceDirect

Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour

The oxidation of organic additives in the positive vanadium electrolyte and its effect on the performance of vanadium redox flow battery Tam D. Nguyen a, b, Adam Whitehead c, Günther G. Scherer d, Nyunt Wai b, Moe O. Oo b, Arjun Bhattarai b, e, Ghimire P. Chandra a, b, Zhichuan J. Xu e, * a

Interdisciplinary Graduate School, Nanyang Technological University, Singapore Energy Research Institute @ Nanyang Technological University, Singapore € GILDEMEISTER Energy Storage GmbH, IZ NO-Süd Straße 3, Objekt M36, AT-2355 Wiener Neudorf, Austria d TUM-CREATE, 1 Create Way, 138602, Singapore e School of Material Science and Engineering, Nanyang Technological University, Singapore b c

h i g h l i g h t s

g r a p h i c a l a b s t r a c t


Organic additives containing polar functional groups could be oxidized by V(V).
The SOC of vanadium electrolyte is reduced to different degrees by the additives.
Some organic additives also affect the performance of a test cell.
A standard screening method for thermally stable additives has been introduced.

a r t i c l e i n f o

a b s t r a c t

Article history: Received 12 August 2016 Received in revised form 23 September 2016 Accepted 5 October 2016 Available online 11 October 2016

Despite many desirable properties, the vanadium redox flow battery is limited, in the maximum operation temperature that can be continuously endured, before precipitation begins in the positive electrolyte. Many additives have been proposed to improve the thermal stability of the charged positive electrolyte. However, we have found that the apparent stability, revealed in laboratory testing, is often simply an artifact of the test method and arises from the oxidation of the additive, with corresponding partial reduction of V(V) to V(IV). This does not improve the stability of the electrolyte in an operating system. Here, we examined the oxidation of some typical organic additives with carboxyl, alcohol, and multi-functional groups, in sulfuric acid solutions containing V(V). The UVevis measurements and titration results showed that many compounds reduced the state-of-charge (SOC) of vanadium electrolyte, for example, by 27.8, 88.5, and 81.9% with the addition of 1%wt of EDTA disodium salt, pyrogallol, and ascorbic acid, respectively. The cell cycling also indicated the effect of organic additives on the cell performance, with significant reduction in the usable charge capacity. In addition, a standard screening method for thermally stable additives was introduced, to quickly screen suitable additives for the positive vanadium electrolyte. © 2016 Elsevier B.V. All rights reserved.

Keywords: Vanadium redox flow battery Organic additive oxidation State-of-charge reduction Cell performance Standard screening method

* Corresponding author. http://dx.doi.org/10.1016/j.jpowsour.2016.10.017 0378-7753/© 2016 Elsevier B.V. All rights reserved.

T.D. Nguyen et al. / Journal of Power Sources 334 (2016) 94e103

1. Introduction The large-scale implementation of renewable energy technologies, such as solar photovoltaic and wind energy, requires effective and stable electric energy storage systems, due to the intermittency of renewable energy. Among some potential candidates, the vanadium redox flow battery (VRFB) has become one of the most promising energy storage systems, thanks to several advantages: power and energy capacity can be independently designed; simple structure of cell and stack design; quick response and long cycle life [1]. Since the pioneering research of M. Skyllas-Kazacos's group in Australia [1], many efforts have been made to overcome the disadvantages of the VRFB. One of these drawbacks is that of the relatively low solubility of V(V) in sulfuric acid [2,3]. The positive electrolyte of the VRFB consists of a mixture of vanadium salts in oxidation states 4þ (IV) and 5þ (V), dissolved in sulfuric acid, with a total vanadium concentration of typically 1.5e2 mol dm3. As the battery is charged, the ratio of V(V) to V(IV) increases. Therefore, the VRFB is generally operated with the positive electrolyte in a metastable state. Consequently, there exists the risk that V(V) ions may condense to form V2O5 and subsequently precipitate, which is almost irreversible in charged electrolyte. In practice, the kinetics of precipitation are very strongly dependent on temperature and state-of-charge (SOC). In sulfuric acid, at typical electrolyte concentrations, V(V) species exist predominantly in the form of the penta-coordinated [VO2(H2O)3]þ cation, at 25  C. At temperatures 40  C precipitation occurs notably, e.g. via the deprotonation and condensation reactions, given below [3]: Deprotonation: 2[VO2(H2O)3]þ  2Hþ / 2VO(OH)3 þ 2H2O Condensation: 2VO(OH)3 / V2O5 þ 3H2O By careful control of temperature and state-of-charge, VRFB manufacturers avoid precipitation. However, these measures add to the cost and can reduce performance. Therefore, much effort has been made to find electrolyte additives that would allow wider operational temperature windows. Chemicals were proposed to be used as thermal stability additives for the positive vanadium electrolyte, based on two main criteria. Firstly, the ability to form soluble neutral species with penta-coordinated V(V) ion, which can prevent deprotonation [4]. Secondly, some organic additives can adsorb, via polar functional groups, such as OH, CHO, C]O, on the initial V2O5 nuclei, hindering further growth and formation of particles large enough to precipitate [5,6]. In the past few years, many organic and inorganic additives have been investigated to improve the stability of the positive vanadium electrolyte at higher temperatures. Many of these organic additives contain one or more polar functional groups. It was reported that the addition of 0.5%wt of some acid compounds, such as methanesulfonic acid, trifluoroacetic acid, polyacrylic acid, oxalic acid, and methacrylic acid could improve the stability of 3 M V(V) in 5 M H2SO4 electrolyte from 5  C to 45  C [7]. Research on 2 M V(V) in 3 M H2SO4 electrolyte showed that the addition of 1%wt L-glutamate could delay the initiation of precipitation in the positive electrolyte for 12 h at 40  C and for 5 h at 50  C [5]. In another work, electrolyte stability tests at 45, 50, and 60  C confirmed that the addition of 0.05e0.1%wt coulter dispersant IIIA (mainly containing coconut oil amine adduct with 15 ethylene oxide groups) to the ~2 M V(V) in 5 M H2SO4 electrolyte, could significantly delay the time of precipitate formation from 1.8e12.3 he30.3 he19.3 days [8]. Polyacrylic acid and its mixture with CH3SO3H were reported as promising stabilizing candidates

95

for vanadium positive electrolyte solution at temperatures higher than 40  C [9]. The addition of 1e4%wt of trishydroxymethylaminomethane (Tris) was also demonstrated to improve the thermal stability of 2 M V(V) in 3 M H2SO4 electrolyte at 40  C [10]. Another study reported that fructose, mannitol, glucose, and D-sorbitol could be used as additives [11]. Inositol, phytic acid and sodium oxalate have been claimed to improve the thermal stability of 1.8 M V(V) in 3 M H2SO4 electrolyte at up to 60  C [6,12]. Several additives, although not tested for thermal stability, were proposed to improve the electrochemical properties of vanadium electrolyte [13]. Besides that, a wide range of organic compounds, proposed by M. Skyllas-Kazacos, have not been thoroughly investigated [1,14]. However, due to the strong oxidative property of V(V) ions, some organic chemicals, especially with the functional groups of OH, CHO, and C]O are known to be unstable in V(V) solutions [7,9]. Waters et al. published a comprehensive study on the oxidation of organic compounds by V(V) in a series of articles in the 1960's, which are especially relevant in light of the most recent VRFB developments [15e28]. It is known that V(V) ions can oxidize some organics in acidic solution and hence be reduced to V(IV), especially over the long testing times typically employed at 40e50  C. However, the effect of organic additives on the positive vanadium electrolyte and especially on the performance of VRFB has not been widely reported. In practice, the rate of precipitation of V2O5 is dependent on the SOC of the electrolyte (i.e. the relative concentration of V(V) to total vanadium). Therefore, it is clear that the apparent thermal stability improvement of an additive may be due to the reduction of V(V) to V(IV) in the test, rather than a genuine hindrance of the condensation and/or polymerization of V(V). To investigate this possibility, we examined the oxidation of some organic compounds in fully charged positive electrolyte. In this study, we classified the examined organic additives into three groups: carboxylic compounds (oxalic acid, sodium oxalate, potassium oxalate and EDTA disodium salt), alcohol compounds (pinacol, resorcinol, methyl resorcinol, xylose, glucose, poly-vinyl-alcohol (PVA) and pyrogallol), and multi-functional group compounds (containing more than 2 different functional groups) (lactic acid, citric acid, ammonium citrate, cysteine, tannic acid and ascorbic acid). The change in the oxidation states of V(V), in the presence of organic additive, was quantified by UVevis spectrometry and also by titration. The influence of organic additives on the performance of the VRFB, was studied by cycling a single cell with 20 cm2 active area. In addition, due to the fact that there was no standard screening procedure for thermally stable additives for vanadium electrolyte, we also introduced a standard additive screening method which could be applied for any proposed additive compound. 2. Experimental method 2.1. Preparation of vanadium positive electrolyte and additives The V(V) solution (100% SOC) was prepared by charging 1.6 M V(III/IV) in 4 M total SO2 4 electrolyte (AMG Titanium Alloys & Coatings, Germany) using a 20 cm2 single VRFB cell. The SOC was determined by measurement of the open circuit voltage (OCV) of the cell. Additives were purchased and divided into three groups: carboxyl compounds (containing only carboxylic group), alcohol compounds (containing only hydroxyl group) and multi-functional group compounds (containing more than 2 different functional groups). The list of the suppliers and grades of the additives are shown in following table:

96

T.D. Nguyen et al. / Journal of Power Sources 334 (2016) 94e103

Carboxyl

Alcohol

Multi-functional group

Oxalic acid (Sigma-Aldrich, purified grade, 99.999%) Sodium oxalate (Sigma-Aldrich, ACS reagent,  99.5%) Potassium oxalate (Sigma-Aldrich, ACS reagent, 99%) EDTA disodium salts (Sigma-Aldrich, 98.5-101.5%)

Resorcinol (Sigma-Aldrich, ACS reagent,  99.0%) Pinacol (Sigma-Aldrich, 98%) 2-Methyl resorcinol (Sigma-Aldrich, 98%) Pyrogallol (Sigma-Aldrich,  98% (HPLC)) Xylose (Alfa-Aesar, 98 þ %) Glucose (Sigma-Aldrich,  99.5%) PVA (Sigma-Aldrich, Mw 9000e10,000, 80% hydrolyzed)

Lactic acid (Sigma-Aldrich,  98%) Cysteine (Sigma-Aldrich, 97%) Ascorbic acid (Sigma-Aldrich, 99%) Citric acid (Sigma-Aldrich, ACS reagent,  99.5%) Ammonium citrate (Sigma-Aldrich, ~98% (capillary GC)) Tannic acid (Sigma-Aldrich, ACS reagent)

2.2. Ultraviolet-visible spectroscopy 1 mL solution of 1.6 M V(V) in 4 M total SO2 4 solution (100% SOC) was prepared with 1.0%wt of selected additives. To achieve a complete dissolution, all samples were sonicated for 1 h and kept for 5 days at room temperature before performing UVevis measurements. A solution of 4 M H2SO4 was used as the reference solution. In practice, an aliquot of 100 mL of sample was diluted to 3 mL with reference solution, to practically eliminate interference from complexes of V(IV) and V(V) [29]. UVeVis spectra were measured on a Shimadzu UV-2501 PC Spectrometer with a 10 mm path-length quartz cell. Spectra of 1%wt of selected additives in 4 M H2SO4 solution (without vanadium), were also recorded, to examine the light absorbance behavior of organic additives in the region of interest. 2.3. Titration Organic compounds were added, to a 1 mL solution of 1.6 M V(V) in 4 M total SO2 4 (100% SOC), to give a concentration of 1%wt and kept at room temperature for 5 days. A solution of 0.02 M KMnO4 (volumetric, Sigma-Aldrich) was used as the oxidizing agent for the titration. All titrations in this work were done by a Mettler Toledo titrator using a DMi144-SC metal sensor (Plug & Play combined platinum ring redox electrode with a ceramic frit for redox titrations with change of the pH value). The titration reaction for V(IV) to V(V) ions by KMnO4 solution can be written as: þ þ 2þ 5VO2þ þ H2 O þ MnO 4 /5VO2 þ 2H þ Mn

A titration for the sample containing 1%wt of selected additives in 4 M H2SO4 solution (without vanadium) was also performed to determine the extent of direct reaction between the organic additives and Mn(VII) ions, over the time period of a typical titration measurement. 2.4. Cell cycling A single cell with 20 cm2 active area was used to perform the cycling test at room temperature (22  C). The main components of the cell included: PVC flow frame, expanded graphite bipolar plate (TF 6, SGL Group), graphite felt porous electrode (GFD 4.6 EA, SGL Group, heat treated at 600  C for 5 h), and anion exchange membrane (FAP 450, Fumatech). The cell was charged and discharged with a current density of 40 mA cm2 and within the potential window of 0.9e1.65 V. 100 mL of vanadium electrolyte was pumped through the cell by a peristaltic pump, with a flow-rate 10 mL/min. Selected organic additives were added to both electrolyte tanks

before the cell cycling test. The charged/discharged cycle was controlled by a NEWARE battery tester.

3. Result and discussion 3.1. Effect of organic additives on the oxidation state of V(V) ions The UVevis spectra of vanadium electrolyte solutions in this study were compared with those reported by Xin [30]. In their research, a distinctive peak was observed for the 100% V(IV) solution at 760 nm, and there was also strong absorption at wavelengths below 340 nm. A 100% V(V) solution was essentially transparent above 500 nm, but absorbed strongly at wavelengths below 500 nm. Fig. 1(aec) show the UVevis spectra of 4 M H2SO4 solution after the addition (1 wt%) of carboxyl compounds (Fig. 1a), alcohol compounds (Fig. 1c), and multi-functional group organic compounds (Fig. 1c), over the wavelength range from 300 to 900 nm. The UVevis spectra of 100% V(IV) and 100% V(V) solutions are also displayed in Fig. 1a for comparison. By adding selected organic compounds, there was no absorption peak over the wavelength range of interest (500e900 nm). Therefore, they are suitable for evaluation by this spectroscopic technique. Fig. 1(def) show the UVevis spectra of 1.6 M V(V) in 4 M total SO2 4 solution after addition of different organic additives, over the wavelength range from 300 to 900 nm. An absorbance peak at 760 nm and an onset of strong absorption below 340 nm is observed in the presence of 1% wt of carboxyl compounds (Fig. 1d), alcohol compounds (Fig. 1e), and multi-functional group organic compounds (Fig. 1f). By comparison with the reference spectra for V(IV) and V(V), it can be seen that V(V) was reduced to V(IV) by the additives listed above. To again confirm the change in the valence state of V(V) from spectroscopic observations, we conducted the titration of V(V) solution containing 1%wt of said organic additives. Fig. 2 aec displays the titration curve of 4 M total SO2 4 solution (without vanadium) with 1%wt of carboxyl compounds (a), alcohol compounds (b), and multi-functional group compounds (c). Compared with the blank 4 M H2SO4 solution curve, it can be seen that resorcinol, methyl resorcinol, pyrogallol, cysteine, tannic acid, and ascorbic acid reacted with Mn(VII) over the timescale of the titration measurement. This could cause inaccuracy in the titration measurement, if the six listed organic additives had not been fully oxidized in V(V) solution after 5 days. However, for the remaining additives the titration method is considered valid. Fig. 2 def show the titration curve of 1.6 M V(V) in 4 M total SO2 4 solution (100% SOC) with 1%wt of selected organic additives. Compared with the reference titration curves of V(IV) and V(V) solution, it was observed that in the presence of every additive, a certain amount of V(V) was reduced to V(IV) and thus the concentration of V(V) decreased.

T.D. Nguyen et al. / Journal of Power Sources 334 (2016) 94e103

97

Fig. 1. UVevis spectra of 4 M H2SO4 solution (aec) and 1.6 M V(V) in 4 M H2SO4 solution (eef) with 1%wt of carboxyl compounds (a, d), alcohol compounds (b, e), and multifunctional group compounds (c, f). The UVevis spectra of V(IV) and V(V) solutions are also displayed for comparison.

3.2. Effect of organic additives on the state-of-charge (SOC) of vanadium electrolyte The SOC of V(V) solution, before and after the addition of different organics additives, was calculated using the Beer-Lambert law, based on the intensity of the absorbance peak at 760 nm [30] and the titration results. Fig. 3 shows the average SOC reduction of V(V) solution (100% SOC) after the addition of 1%wt of carboxyl compounds (a), alcohol compounds (b), and multi-functional group organic compounds (c). It can be noted that all of these organic additives reduced the SOC of the positive vanadium electrolyte under static conditions. In detail, for carboxyl additives, 1%wt of EDTA disodium salts could reduce up to 27.8% of the SOC of original V(V) solution, while the value for oxalic acid, potassium oxalate and sodium oxalate are 12.5%, 8.9% and 8.4%, respectively. For alcohol compounds, 1%wt of pyrogallol and glucose could diminish the SOC by about 88.5% and 77.4%, correspondingly. Xylose and resorcinol were similar, at about 68%, and pinacol showed the lowest reduction of SOC, at about 32.7%. For multi-functional group additives,

ascorbic acid with 1% wt concentration could cut down 81.9% SOC of V(V) solution. Tannic acid and citric acid decreased the SOC by about 70.3 and 63.7%. The lowest reduction was observed when adding 1%wt of lactic acid, at about 38.2% of SOC. To clarify the reactivity of each organic compound, we calculated the associated number of electrons involved in oxidizing the additive, which is indicated in Fig. 4. It is known that the vanadium reduction from V(V) to V(IV) goes through a one electron reaction: 2þ þ  VOþ þ H2 O 2 þ 2H þ e 4VO

Combined with the change in the SOC of V(V) solution, we can calculate the total number of electrons involved in the oxidation reaction of additives, and from that, we can get the average number of electron involved per molecule of additive (the additive concentration was calculated by conversion of weight percentage to molar percentage). In general, from the result displayed in Fig. 4, we observed that the extent of additive oxidation increased with the increasing number of functional groups in the additive's molecular

98

T.D. Nguyen et al. / Journal of Power Sources 334 (2016) 94e103

Fig. 2. Titration curve of 4 M H2SO4 solution (aec) and 1.6 M V(V) in 4 M H2SO4 solution (eef) with 1%wt of carboxyl compounds (a, d), alcohol compounds (b, e), and multifunctional group compounds (c, f).

structure. In detail, for carboxyl additives, the most extensively oxidized additive was EDTA disodium salt, which underwent an oxidation reaction involving ~12 e/molecule. EDTA possesses 4 carboxylic groups compared to two carboxylic groups for oxalic acid and its salts. However, the latter only underwent oxidation by about 1.3 e/molecule after 5 days at room temperature. For alcohol additives, the most extensively reductive additive was glucose, which released about 16.3 ee/molecule, while the weakest additive was pinacol, which produced about 4.5 ee/molecule during oxidation. For multi-functional group compounds, we recorded the number of 16.8 ee/molecule for ascorbic acid, which denotes as the strongest reductive agent, while lactic acid was the weakest reductive chemical, which showed the value of 4 ee/molecule during oxidation. Two high molecule weight compounds not listed in Fig. 4, tannic acid and PVA, revealed a high value of oxidation degree, up to 139.5 and 808 ee/molecule, respectively. Converted to their corresponding monomers (as-studied tannic acid contains 10 molecules of gallic acid, while PVA contains 216 monomers of vinyl

alcohol), the recorded values are ~13.95 and ~3.74 ee/molecule for gallic acid and vinyl alcohol, correspondingly. Selected examples will be examined in more detail. Oxalic acid can theoretically undergo complete oxidation through a 2e process in aqueous acid solution [22,31]:

T.D. Nguyen et al. / Journal of Power Sources 334 (2016) 94e103

Fig. 3. The state-of-charge (SOC) reduction after 5 d at room temperature of 100% SOC vanadium electrolyte after the addition of 1%wt of carboxyl compounds (a), alcohol compounds (b), and multi-functional group compounds (c). All the values are the average calculation from titration results and estimation by Beer-Lambert relationship.

Oxidation : HOOCCOOH42Hþ þ 2CO2 þ 2e 2þ þ  Reduction : 2VOþ þ 2H2 O 2 þ 4H þ 2e 42VO þ þ Overall : HOOCCOOH þ 2VO2 þ 2H 42VO2þ þ 2H2 O þ 2CO2

99

Fig. 4. The degree of additive oxidation in V(V) solution of carboxyl compounds (a), alcohol compounds (b), and multi-functional group compounds (c).

In practice, we observed the number of 1.3 e/molecule (Fig. 4), which implies that the reaction goes about 65% through to completion. One study reported that the oxidation of lactic acid by V(V) ions theoretically follows a 2e process to form acetaldehyde in sulfuric acid medium at 30  C [32]:

100

T.D. Nguyen et al. / Journal of Power Sources 334 (2016) 94e103

The oxidation process of other organic compounds is complicated. Analysis of the decomposition products between the organic molecules and V(V) in sulfuric acid would be required to support mechanistic arguments. However, what is obvious from these findings is that the charged positive vanadium electrolyte is a strongly oxidizing media, where all of the proposed additives were, to a greater or lesser extent, reacted.

3.3. Effect of organic additives on the cell performance

Oxidation: H3CCHOHCOOH 4 CH3CHO þ CO2 þ 2Hþ þ 2ee However, in practice we recorded the number of ~ 4e/molecule (Fig. 4), which is most likely to be resulting from oxidation to acetic acid:

Under static conditions, we have observed the effect of selected organic additives on the valence state of V(V), and thus on the SOC of positive vanadium electrolyte. However, it is also very necessary to evaluate their influence in practical cell operation. Fig. 5 indicates the coulombic efficiency (CE), energy efficiency (EE) and voltage efficiency (VE) of a 20 cm2 active area single cell over 11 cycles, with the addition of three of the strongest reducing agents from each of the groups of selected additives: EDTA disodium salts,

Oxidation : H3 CCHOHCOOH þ H2 O4CH3 COOH þ CO2 þ 4Hþ þ 4e 2þ þ  Reduction : 4VOþ þ 4H2 O 2 þ 8H þ 4e 44VO þ Overall : H3 CCHOHCOOH þ 4VO2 þ 4Hþ 4CH3 COOH þ 4VO2þ þ 3H2 O þ CO2

For aldose sugars such as glucose, a study in acid medium claimed that the oxidation of glucose by V(V) involved only 2ee in 2 M H2SO4 (or HClO4) solution at 40e50  C [33]:

C6H12O6 þ 2V(V) þ H2O 4 C5H10O5 þ HCOOH þ 2V(IV) þ 2Hþ In another paper using perchloric acid at 45  C, the oxidation of glucose by V(V) was more intensive, at 12ee/molecule [34]: C6H12O6 þ 12V(V) þ 6H2O 4 6HCOOH þ 12V(IV) þ 12Hþ However, in our case, the oxidation of glucose by V(V) in sulfuric acid proceeded through ~16 ee/molecule (Fig. 4). To satisfy this number, the reaction has to form 4 equivalents of formic acid and 2 of CO2, for example:

glucose and ascorbic acid. By adding 1%wt of EDTA disodium salts, which underwent oxidation reactions involving ~12 e/molecule and could reduce about 27.8% SOC of full charged positive vanadium electrolyte under static conditions in vitro; during cycling, the CE, EE and VE of the cell were lower, on the first cycle, than when using blank electrolyte. This could be because oxidation of the EDTA disodium salts occurred predominately during the first charge/ discharge cycle. For the addition of 1%wt of glucose, which underwent an oxidation reaction involving ~16 e/molecule and reduced the positive electrolyte SOC by 77.4% under static conditions, the CE, EE and VE of the cell fluctuated erratically for the first 6 cycles. They stabilized thereafter, but at slightly lower values than for the blank electrolyte. This may be due to the slow kinetics of the oxidizing reaction, which needed 6 cycles to completely oxidize the glucose. The erratic cycling is also probably due to CO2 gas bubbles in the cell, which are formed as a byproduct of the glucose oxidation. Similar to glucose, the addition of 1%wt of ascorbic acid, which underwent an oxidation reaction involving ~16.8 e/molecule and reduced the positive electrolyte SOC by 81.9% under static conditions, the CE, EE and VE of the cell also oscillated unpredictably for the first 6 cycles. This may be also due to the slow kinetics of the oxidizing reaction, which needed 6 cycles to completely oxidize the ascorbic acid. The formation of CO2 gas bubbles in the cell, which are formed as a byproduct of the ascorbic acid oxidation, also may cause to the unpredicted cycling. The influence of those organic additives on the cell performance is more clearly seen when investigating the capacity. Fig. 6 indicates the capacity drop of the cell during 11 charged/discharged cycles (a)

Oxidation : C6 H12 O6 þ 6H2 O44HCOOH þ 2CO2 þ 16Hþ þ 16e 2þ þ  Reduction : 16VOþ þ 16H2 O 2 þ 32H þ 16e 416VO þ þ Overall : C6 H12 O6 þ 16VO2 þ 16H 416VO2þ þ 4HCOOH þ 2CO2 þ 10H2 O

T.D. Nguyen et al. / Journal of Power Sources 334 (2016) 94e103

101

Fig. 6. (a) Capacity drop of VRFB during 11 charged/discharged cycles with the addition of 1%wt EDTA disodium salts, glucose, and ascorbic acid. (b) Capacity drop vs. cell voltage after 10 cycle of VRFB with the addition of 1%wt EDTA disodium salts, glucose, and ascorbic acid.

the addition of EDTA disodium salt could reduce about 10.5% of cell charge capacity, from ~3.0 Ah to 2.7 Ah, the addition of glucose and ascorbic acid caused a more significant reduction, about 59.7% of cell capacity for both chemicals. The greater reduction in usable charge capacity after 10 cycles is expected from the static experiments, and confirms that oxidation of the organic additives also occurs under charge/discharge cycling of the VRFB. In practical operation, all organic additives mentioned in this study will not only reduce the performance of VRFB, their apparent thermal stabilizing ability is due, probably to a greater extent, to the reduction of V(V). 3.4. Standard screening method for vanadium thermal stability additives Fig. 5. Efficiency of VRFB with the addition of 1%wt EDTA disodium salts, glucose, and ascorbic acid.

and capacity drop vs. cell voltage after 10 cycles (b) when adding 1% wt of EDTA disodium salts, glucose, and ascorbic acid. From Fig. 6 (a), it can be seen that by adding these three selected organic additives, the capacity of the cell dropped rapidly, especially for the addition of glucose and ascorbic acid. In detail, in Fig. 6 (b), while

It is known that the screening of thermally stable additives for the vanadium electrolyte requires a long testing time, due to the slow kinetics of V2O5 precipitation at the temperatures of 40e50  C. This effort is obviously wasted if the additive is unstable. Therefore, a quick method to determine and eliminate unstable additives from lengthy test programs is highly desired. In Fig. 7, we proposed a standard screening method for vanadium thermal stability additives. The proposed chemicals have to undergo three consecutive

102

T.D. Nguyen et al. / Journal of Power Sources 334 (2016) 94e103

References

Fig. 7. Standard screening method for thermal stable additives for positive vanadium electrolyte.

experiments to demonstrate their chemical stability and thermal stability properties. Firstly, UVevis spectra can quickly indicate the change in the oxidation state of V(V) ions in the presence of additive. However, it could happen that the additives absorb over the same wavelength range as the V(IV) solution. The second test, titration against Mn(VII), is optional, but strongly advised in instances where the additive is colored. It can provide similar information to the spectroscopy result, assuming that the reaction of the additive is not very rapid with Mn(VII). Finally, if a chemical appears stable to the oxidation by V(V) ions, it now can be used for the thermal stability test. This is carried out with fully charged electrolyte, in glass test tubes, under thermostated conditions at 40  Ce50  C, and compared against an additive-free “blank” solution. 4. Conclusion In this study, we have demonstrated the oxidation of a range of organic additive compounds in positive vanadium electrolyte. We suggested that many of the claims regarding improved thermal stability of positive electrolyte with these additives are, to some extent, artifacts, caused by reduction of the SOC. All of the selected additives reduced the SOC of charged vanadium electrolyte, and would therefore have an influence on the performance of vanadium redox flow battery. The proposed interaction between V(V) ions and some organic compounds was also explained by the degree of additive oxidation. In addition, a standard screening method for positive vanadium electrolyte additives was also introduced to reduce the testing time. This study will provides insight into the effect of many organic additives on the vanadium redox flow battery electrolyte. Acknowledgement The present work was supported by Energy Research Institute @ Nanyang Technological University, Gildemeister Energy Storage and the SGL Group.

[1] M. Skyllas-Kazacos, M. Kazacos. High energy density vanadium electrolyte solutions, methods of preparation thereof and all-vanadium redox cells and batteries containing high energy vanadium electrolyte solutions. Patent U. S. 7078123 B2, 2006. [2] D.-G. Oei, Chemically regenerative redox fuel cells, J. Appl. Electrochem. 12 (2012) 41e51. [3] M. Vijayakumar, L. Li, G. Graff, J. Liu, H. Zhang, Z. Yang, J.Z. Hu, Towards understanding the poor thermal stability of V5þ electrolyte solution in Vanadium Redox Flow Batteries, J. Power Sources 196 (2011) 3669e3672. [4] L. Li, S. Kim, W. Wang, M. Vijayakumar, Z. Nie, B. Chen, J. Zhang, G. Xia, J. Hu, G. Graff, J. Liu, Z. Yang, A stable vanadium redox-flow battery with high energy density for large-scale energy storage, Adv. Energy Mater. 1 (2011) 394e400. [5] Y. Lei, S. Liu, C. Gao, X. Liang, Z. He, Y. Deng, Z. He, Effect of amino acid additives on the positive electrolyte of vanadium redox flow batteries, J. Electrochem. Soc. 160 (4) (2013) A722eA727. [6] X. Wu, S. Liu, N. Wang, S. Peng, Z. He, Influence of organic additives on electrochemical properties of the positive electrolyte for all-vanadium redox flow battery, Electrochim. Acta 78 (2012) 475e482. [7] G. Wang, J. Chen, X. Wang, J. Tian, H. Kang, X. Zhu, Y. Zhang, X. Liu, R. Wang, Study on stabilities and electrochemical behavior of V(V) electrolyte with acid additives for vanadium redox flow battery, J. Energy Chem. 23 (2014) 73e81. [8] F. Chang, C. Hua, X. Liu, L. Liu, J. Zhang, Coulter dispersant as positive electrolyte additive for the vanadium redox flow battery, Electrochim. Acta 60 (2012) 334e338. [9] J. Zhang, L. Li, Z. Nie, B. Chen, M. Vijayakumar, S. Kim, W. Wang, B. Schwenzer, J. Liu, Z. Yang, Effects of additives on the stability of electrolytes for allvanadium redox flow batteries, J. Appl. Electrochem. 41 (2011) 1215e1221. [10] S. Peng, N. Wang, C. Gao, Y. Lei, X. Liang, S. Liu, Y. Liu, Influence of trishydroxymethyl aminomethane as a positive electrolyte additive on performance of vanadium redox flow battery, Int. J. Electrochem. Sci. 7 (2012) 2440e2447. [11] S. Li, K. Huang, S. Liu, D. Fang, X. Wu, D. Lu, T. Wu, Effect of organic additives on positive electrolyte for vanadium redox battery, Electrochim. Acta 56 (2011) 5483e5487. [12] D. Kim, J. Jeon, An electrolyte with high thermal stability for the vanadium redox flow battery, Adv. Mater. Mech. Struct. Eng. (2016) 371e375. Taylor &Fracis Group, London, ISBN: 978-1-138-02908-8. [13] Z. Jia, B. Wang, S. Song, X. Chen, Effect of polyhydroxy-alcohol on the electrochemical behavior of the positive electrolyte for vanadium redox flow batteries, J. Electrochem. Soc. 159 (6) (2012) A843eA847. [14] M. Skyllas-Kazacos, M. Kazacos. Stabilised electrolyte solutions, methods of preparation thereof and redox cells and batteries containing stabilised electrolyte solutions. Patent WO 1995012219 A1, 1995. [15] J.S. Littler, W.A. Waters, Oxidations of organic compounds with quinquevalent vanadium. Part I. General survey; and the oxidation of pinacol, J. Chem. Soc. (1959) 1299e1305. [16] J.S. Littler, W.A. Waters, Oxidations of organic compounds with quinquevalent vanadium. Part II. The oxidation of ketones, J. Chem. Soc. (1959) 3014e3019. [17] J.S. Littler, W.A. Waters, Oxidations of organic compounds with quinquevalent vanadium. Part III. The oxidation of cyclohexanol, J. Chem. Soc. (1959) 4046e4052. [18] J.S. Littler, A.I. Mallet, W.A. Waters, Oxidations of organic compounds with quinquevalent vanadium. Part IV. The oxidation of some a-glycols, J. Chem. Soc. (1960) 2761e2766. [19] J.S. Littler, W.A. Waters, Oxidations of organic compounds with quinquevalent vanadium. Part V. A comparative study of the oxidation of alcohols and glycols by cerium (IV), vanadium(V), and chromium (VI), J. Chem. Soc. (1960) 2767e2772. [20] J.R. Jones, W.A. Waters, Oxidations of organic compounds with quinquevalent vanadium. Part VI. Carbon-carbon bond fission in the oxidations of 2-phenylethanol and a-t-butylbenzyl alcohol, J. Chem. Soc. (1960) 2772e2784. [21] J.R. Jones, W.A. Waters, J.S. Littler, Oxidations of organic compounds with quinquevalent vanadium. Part VII. The kinetic resemblance between the oxidations of some a-hydroxy-acids and that of pinacol and of its monomethyl ether, J. Chem. Soc. (1961) 630e633. [22] J.R. Jones, W.A. Waters, Oxidations of organic compounds with quinquevalent vanadium. Part VIII. The oxidation of oxalic acid, J. Chem. Soc. (1961) 4757e4761. [23] K. Julian, W.A. Waters, Oxidations of organic compounds with quinquevalent vanadium. Part IX. Bromide-catalysis of some reactions, J. Chem. Soc. (1962) 818e821. [24] J.R. Jones, W.A. Waters, Oxidations of organic compounds with quinquevalent vanadium. Part X. The oxidations of acetoin and 3-hydroxy-3-methylbutan-2one, J. Chem. Soc. (1962) 1629e1631. [25] B.W. Bycroft, J.C. Roberts, Oxidations of organic compounds with quinquevalent vanadium. Part XI. Relative rates of alcohol oxidation, J. Chem. Soc. (1962) 2063e2084. [26] J.R. Jones, W.A. Waters, Oxidations of organic compounds by quinquevalent vanadium. Part XII. Oxidation of propionaldehyde, n-butyraldehyde, and isobutyraldehyde, J. Chem. Soc. (1963) 352e356. [27] T.J. Kemp, W.A. Waters, Oxidations of organic compounds with quinquevalent vanadium. Part XIII. The initial rates of the oxidations of formic and malonic acids, J. Chem. Soc. (1964) 1610e1614.

T.D. Nguyen et al. / Journal of Power Sources 334 (2016) 94e103 [28] T.J. Kemp, W.A. Waters, Oxidations of organic compounds with quinquevalent vanadium. Part XIV. Catalysis of the oxidations of malonic and oxalic acids by manganous ions, J. Chem. Soc. (1964) 3101e3106. [29] A.H. Jassim, B.B. Qassim, A.S. Mansoor, Removal interferences with spectrophotometric study for the determination of chromium, vanadium and their application, IBN Al-Haitham J. Pure Appl. Sci. 23 (2010) 3. [30] X. Gao, R.P. Lynch, M.J. Leahy, D.N. Buckley, Spectroscopic study of vanadium electrolytes in vanadium redox flow battery (VRFB), ECS Trans. 45 (26) (2013) 25e36. re, L.A.G. Rodenas, P.J. Morando, M.A. Blesa, Reduction of [31] V.I.E. Bruye

103

vanadium(V) by oxalic acid in aqueous acid solutions, J. Chem. Soc. Dalton Trans. (2001) 3593e3597. [32] R.M. Clementin, K. Takashima, Oxidation of lactic acid by vanadium (V) in sulphuric acid medium: a kinetic and mechanistic study, Quim. Nova 16 (6) (1993). [33] A. Kumar, R.N. Mehrotra, Kinetics of oxidation of aldo sugars by quinquevalent vanadium ion in acid medium, J. Org. Chem. 40 (9) (1975) 1248e1252. [34] K.K.S. Gupta, S.N. Basu, Kinetics and mechanism of oxidation of d-glucose and d-ribose by chromium(VI) and vanadium(V) in perchloric acid medium, Carbohydr. Res. 80 (1980) 223e232.