- Email: [email protected]
Detection of capacity imbalance in vanadium electrolyte and its electrochemical regeneration for all-vanadium redox-flow batteries
Journal of Power Sources 302 (2016) 79e83
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Short communication
Detection of capacity imbalance in vanadium electrolyte and its electrochemical regeneration for all-vanadium redox-flow batteries Nataliya Roznyatovskaya*, Tatjana Herr, Michael Küttinger, Matthias Fühl, Jens Noack, Karsten Pinkwart, Jens Tübke Fraunhofer Institute for Chemical Technology, Joseph-von-Fraunhofer-Str. 7, Pfinztal 76327, Germany
h i g h l i g h t s UVevis detection of V(III) fraction all-vanadium redox-flow battery electrolyte. Isosbestic point at 600 nm as a reference point for common electrolyte batches. Rebalancing procedure based on pre-charging and remixing.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 19 August 2015 Received in revised form 18 September 2015 Accepted 8 October 2015 Available online xxx
A vanadium electrolyte for redox-flow batteries (VRFB) with different VIII and VIV mole fractions has been studied by UVevis spectroscopy. Spectrophotometric detection enables a rough estimate of the VIV and VIII content, which can be used to detect an electrolyte capacity imbalance, i.e. a deviation in the mole fraction of VIV or VIII away from 50%. The isosbestic point at 600 nm can be used as a reference point in the analysis of common VRFB electrolyte batches. The VRFB electrolyte is observed to have an imbalance after prolonged storage (a couple of years) in a tank under ambient conditions. A regeneration procedure, which involves pre-charging the unbalanced electrolyte and mixing part of it with a portion of initial unbalanced electrolyte, has been tested. The resulting rebalanced electrolyte has been compared with a common electrolyte in a chargeedischarge cell test and is shown to be suitable for cell operation. © 2015 Elsevier B.V. All rights reserved.
Keywords: Vanadium redox-flow battery UVevis absorption Unbalanced state-of-charge Electrolyte
1. Introduction Vanadium salts dissolved in sulfuric acid are used for energy storage in all-vanadium redox-flow batteries (VRFBs) [1,2]. During battery discharge vanadium species undergo redox reactions of VV to VIV in the positive half-cell, and VII to VIII in the negative half-cell. A mixture of VIII and VIV species in a 50:50% mole ratio (referred to as V3.5þ electrolyte) is often used as a commercially-available starting electrolyte for VRFBs, and needs to be charged to form VIII as an anolyte and VIV as a catholyte for further battery operation. During long-term performance the VRFB electrolyte is reported to become unbalanced [3e5]. Imbalance implies an inequality in the amount of oxidized and reduced species in a VRFB, and is usually caused either by vanadium cross-over through the
* Corresponding author. E-mail address: [email protected] (N. Roznyatovskaya). http://dx.doi.org/10.1016/j.jpowsour.2015.10.021 0378-7753/© 2015 Elsevier B.V. All rights reserved.
membrane or by side reactions such as hydrogen evolution in the anolyte. It results in capacity losses, and 100% state-of-charge (SoC) can then be achieved only in one half-cell of the VRFB. To detect the imbalance, the oxidation state and the concentration of vanadium species can be determined and compared. A direct analysis can be performed based on the potentiometric titration of VRFB samples; however, this is time-consuming and unsuitable for continuous electrolyte monitoring or for monitoring online. Indirectly, the SoC can be estimated from its correlation with the electrolyte absorption in the visible range (although this is difficult to use in practice for the catholyte [6]) or from the correlation between the open circuit potential and the vanadium concentration. For the latter technique, a special reference cell is connected to the electrolyte circuit, and the electrical measurement set-up is adapted to it [4,5]. All the above mentioned approaches to detect vanadium electrolyte imbalance concern the analysis of the anolyte and catholyte, i.e. mixtures of VII/VIII and VIV/VV. In this paper we report the imbalance of common V3.5þ electrolyte, which is found to occur during its long-term storage before it is introduced into the battery.
80
N. Roznyatovskaya et al. / Journal of Power Sources 302 (2016) 79e83
As a powerful VRFB system (for example 2 MW/20 MWh) requires several hundred m3 of electrolyte, it is necessary to transport and store batches of electrolyte and to be able to monitor the consistency of their chemical properties over time before they are introduced into the battery. This electrolyte undergoes an alteration of VIII content during storage in a tank over a couple of years, and the ratio of VIII to VIV consequently shifts. As vanadium electrolyte is an important factor in the total costs of the VRFB system, and the published literature on electrolyte rebalancing mostly considers the concentration shifts in the electrolyte during battery operation, we here present an approach for the rapid spectrophotometric identification of imbalance in common V3.5þ electrolyte, and a procedure for the electrolyte pre-treatment to overcome the imbalance. The proposed pre-treatment is based on electrolysis and mixing of electrolyte portions to obtain separate VIII and VIV electrolyte solutions. 2. Experimental 2.1. Chemicals and material Vanadium electrolyte solution (V3.5þ) with a 50:50% mole fraction of VIII and VIV at a total vanadium concentration of 1.55 M, 2 M sulfuric acid and approx. 0.05 M phosphoric acid (GfE Gesellschaft für Elektrometallurgie mbH, Germany) was used as a reference electrolyte, and to prepare samples with various VIII:VIV ratios. Samples of the same V3.5þ electrolyte which had been stored in a tank and exposed to ambient temperature fluctuations over a couple of years were used as an unbalanced electrolyte. The Fumasep FAP-450 membrane (FuMa-Tech GmbH, Germany) was used as the separator. GFA6 graphite felt (SGL Group, Germany) and FU 4369 graphite bipolar plates (Schunk Kohlenstofftechnik GmbH, Germany) were used as electrode current collector materials in a 40 cm2 cell assembly for the chargeedischarge test (for details of cell design, see Ref. [6]). The felts were pre-treated for one hour at 400 C in the oven before the cell assembly, and were compressed by 10% of their thickness in the cell. The electrolyte flow rate was 75e80 ml/min. A 40 cm2 cell of the same design was used for galvanostatic electrolysis to prepare the electrolyte samples with various VIII:VIV ratios and for pre-charging (i.e. rebalancing) of the unbalanced electrolyte. The vanadium concentrations in all the electrolyte samples were determined by potentiometric titration. 2.2. UVevis characterization
out with 0.1 M cerium(IV) sulfate standard solution using Titrator T70 (Mettler Toledo Int. Inc., Germany).
3. Results and discussion 3.1. Visible spectroscopy characterization To model the unbalanced electrolyte samples with different formal charges from 3þ to 4þ but a constant total vanadium concentration, the common V3.5þ electrolyte was galvanostatically charged over a predefined time. These samples were studied using UVevis spectroscopy. The absorbance spectra of these electrolyte solutions with different ratios of VIII to VIV are shown in Fig. 1. The absorption spectrum of neat VIII has two characteristic peaks at 401 and 600 nm. The extinction coefficients of VIII estimated in the concentration range from 0.2 M to 1.6 M are 13.1 M1 cm1 at 401 nm and 7.5 M1 cm1 at 600 nm. The absorption spectrum of neat VIV shows a band positioned at a maximum of 760 nm with a shoulder at 600 nm [7,8]. A linear dependence between the absorption at 760 nm and the VIV concentration has been observed, at least in the range from 0.2 M to 1.6 M. The extinction coefficient of VIV at 760 nm is 19.5 M1 cm1. It can be seen that the spectra of mixtures consisting of VIII and VIV are the convolution of VIII and VIV absorptions. The evolution of spectra with an increase in the VIV fraction results in the appearance of an isosbestic point, which is consistent with the fact that the common electrolyte is a mixture of non-interacting VIII and VIV species and that the analytical concentration of vanadium is constant throughout the series [6]. All three characteristic peaks can be seen for the 50:50% mixture of VIII and VIV. The unbalanced electrolyte displays a lower absorption at the isosbestic point (dashed line, Fig. 1), so the total vanadium concentration in this sample is lower. This is in good agreement with redox titration data: the total vanadium concentration is 1.4 M compared to an initial concentration of 1.55 M. It means that the slow oxidation of VIII by air cannot be the main reason for the shift in the VIII:VIV ratio, and VIII is likely to precipitate out of the electrolyte solution. To be used for quantitative analysis, VIII and VIV bands, which are overlapped on the spectra of the mixtures (curves (2) to (10) in Fig. 1), should be deconvoluted, especially for VIV fractions up to 40%. However, as can be seen in Fig. 2, the ratio of absorbance taken at 760 nm and at 401 nm (Abs760/Abs401) can be used to roughly detect an VIII:VIV ratio shift. The dependence of Abs760/Abs401 in
The absorbance of vanadium electrolyte samples was recorded with a Shimadzu 1650PC UVevis double beam spectrophotometer. These measurements were carried out in matched quartz cuvettes with a 0.2 mm optical path length. 2.3. Electrochemical procedures The charging of electrolyte for rebalancing was carried out galvanostatically at 20 mA cm2 until the redox potential of the anolyte reached approx. 0.5 V vs. Hg/Hg2SO4 electrode. The redox potential of the anolyte was monitored using a glassy carbon electrode incorporated into the electrolyte circuit. The chargeedischarge cell test was performed using a Basytec battery tester (Basytec GmbH, Germany) in galvanostatic mode: the end-ofcharge voltage was set to 1.65 V and the end-of-discharge voltage was 0.8 V. Between the cycles the cell was kept at open circuit potential for five minutes. At each current density (from 25 mA cm2 to 100 mA cm2) five cycles of chargeedischarge were performed. Each chargeedischarge measurement was repeated for reproducibility. The titration of electrolytes samples was carried
Fig. 1. Evolution of absorption spectra in the visible range for vanadium electrolyte as a function of the VIV molar fraction. Total vanadium concentration is 1.55 M for samples (1)e(11), and 1.4 M for the sample (12).
N. Roznyatovskaya et al. / Journal of Power Sources 302 (2016) 79e83
81
the concentration of VIV (65%). This is in good agreement with the titration data, which indicated that the altered electrolyte has the formal vanadium oxidation state þ3.65. Furthermore the spectra of vanadium electrolytes with different VIII:VIV ratios in the range from 0 to 100% can therefore be used for a qualitative estimation of changes in the total concentration of vanadium.
3.2. Electrochemical treatment
Fig. 2. Correlation of the ratio of absorptions at 760 nm and 401 nm, corresponding to characteristic peaks of VIII and VIV, to VIV content.
logarithmic scale on the percentage of VIV is linear over the range from 40% to 90% of VIV, and the Abs760/Abs401 ratio for the unbalanced electrolyte calculated from measured UVevis spectrum (Fig. 1, curve (12)) and plotted in this linear region corresponds to
Generally there are several approaches to restore the VIII:VIV ratio: (i) Adding VIII salt to compensate the loss or decrease in its concentration; this is, however, expensive (VIII is available mainly as vanadium(III) oxide). (ii) Reducing an excess of VIV chemically, i.e. introducing foreign species into the electrolyte (the electrolyte is contaminated by the products of this reaction). (iii) Electrochemical oxidation or reduction of the electrolyte. The electroreduction of an unbalanced electrolyte to V3.5þ can be performed in an anodic halfcell coupled with an oxygen evolution reaction from sulfuric acid solution in the cathodic half-cell. Here special care should be taken in the catalysis of the oxygen evolution reaction. Furthermore it is not always straightforward to set the end point of the electrolysis and achieve the predefined VIII:VIV ratio. In a rebalancing method proposed in Ref. [3] a lack or excess of charge during electrolysis of
Fig. 3. Schematic representation of electrolyte rebalancing procedure. Colors (green, blue, yellow) correspond to the vanadium species in oxidation states þ3, þ4 and þ5. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
82
N. Roznyatovskaya et al. / Journal of Power Sources 302 (2016) 79e83
Table 1 The quantitative composition of vanadium electrolyte samples in the course of pretreatment for rebalancing:a e determined by titration,b e expected, CV e total vanadium concentration. Sample
Anolyte III
Catholyte
V , % V , % CV, M VIII, % VIV, % VV, % CV, M (1) Unbalanced electrolyte 35a (2) After charging 99a 100b (3) After mixing
IV
65a 1a 0b
1.43a 1.46a
35a
1a 0b
65a 75a 70b 99a 100b
25a 30b
1.43a 1.43a 1.44a
the unbalanced electrolyte is compensated by hydrogen or oxygen evolution reactions (electrolysis of water) following the electrochemical conversion of vanadium, to obtain an electrolyte with SoC either 0% or 100%. In this case an addition of water is needed and special oxidation resistant electrodes for cathodic half-cells were used. In this study the pre-treatment of an unbalanced electrolyte comprised three steps: (i) measurement of the VIII:VIV ratio, (ii) initial charging of the electrolyte to generate a 100% mole fraction of VIII in the anodic compartment, whereby a VIV/VV mixture is formed in the cathodic compartment (iii) mixing of the VIV/VV electrolyte with an initial electrolyte sample in a volumetric ratio calculated to ensure that all the VV reacts with VIII to form VIV. This ratio is calculated based on reaction stoichiometry and concentrations of vanadium species in the initial and electrolyzed samples. The procedure is shown in Fig. 3, in which the colors correspond to the vanadium species in different oxidation states presenting in solution. This procedure results in an excessive amount of electrolyte in the form of VIV, which cannot be directly used as an electrolyte, but no additional cell arrangement or special electrodes are needed. The residual amount of VIV catholyte can in principle be used for vanadium recycling. The total vanadium concentration remains the same as in the initial unbalanced electrolyte sample, because this work was focused on rebalancing of the VIII:VIV ratio. Each step of the electrolyte pretreatment was followed by titration to monitor the redox conversion of vanadium species (Table 1). It can be seen that unbalanced anolyte (V3.65þ) is reduced to VIII. The increase in the total vanadium concentration by 2% of its initial value is unexpected. This is probably caused by water diffusion through the membrane, resulting in anolyte volume
change. The catholyte is oxidized to the formal state V4.25þ (75% of VIV and 25% of VV), which is slightly lower than the expected V4.3þ. The reason for this small deviation is unclear. The catholyte is then mixed with initial unbalanced electrolyte in a volume ratio calculated to ensure that all the VV reacts with VIII to form VIV. For the unbalanced electrolyte in this work the ratio was 1:1 v/v. The formal charge of vanadium electrolyte decreased to þ3.99 after the mixing. Both the catholyte and anolyte generated were introduced into the cell for the chargeedischarge test. 3.3. Cell test study As the proton and counter-anion balance is not usually considered during the charging the electrolyte, it is important to know whether the rebalanced electrolyte can be used for cell operation at the same efficiency as the initial V3.5þ electrolyte. Chargeedischarge cell tests were therefore performed with the electrolyte after capacity rebalancing. For comparison the same tests were performed with a cell filled with common (V3.5þ) batch electrolyte taken as a reference. Energy efficiency, coulomb efficiency and discharge power density were evaluated in these tests, and the results are shown in Fig. 4. The cell operated with V3.5þ electrolyte can be discharged at a maximum current density of 50 mA cm2, attaining a discharge power density of 45 mW cm2. The maximum energy efficiency is 76% at 25 mA cm2. From the results shown in Fig. 4 it can be concluded that the cell operated with rebalanced electrolyte demonstrates the same parameters as the cell with standard electrolyte, even though the total vanadium concentration in the rebalanced electrolyte is 1.4 M compared to 1.6 M in the common electrolyte. The difference of 12% in the total vanadium concentration does not, therefore, have a significant influence on the cell performance. 4. Conclusions Vanadium electrolyte consisting of a mixture of VIII and VIV salts in sulfuric acid is found to undergo aging, i.e. an alteration of the VIII mole fraction and total vanadium concentration during storage for a couple of years under ambient conditions, i.e. without maintaining stable temperature and protection against oxygen. This alteration can be detected by spectrophotometry in the visible range. The ratio of adsorption peaks of VIII and VIV in logarithmic scale is a linear function of the VIV mole fraction, at least if VIV exceeds 40%. This ratio can be used to roughly estimate the VIV and VIII content. A rebalancing process for the unbalanced electrolyte has been proposed and examined. The compensation of accumulated VV in the catholyte, after charging the electrolyte to obtain SoC 0%, is carried out by mixing the generated catholyte with initial unbalanced electrolyte. VV then reacts chemically with VIII to form VIV. The mixing protocol can be adjusted for defined VIII:VIV ratios in unbalanced electrolyte. Acknowledgments This work is financially supported by the State of BadenWuerttemberg and the German Federal Ministry of Education and Research (BMBF) in the context of the project “RedoxWind”. The authors are grateful to Mr. Colin Dessornes and Ms. Jessica Kunzendorf for technical assistance in this work. References
Fig. 4. Parameters evaluated from the chargeedischarge tests on the redox-flow battery with common Vþ3.5 electrolyte (solid lines) and capacity rebalanced electrolyte (dashed lines).
[1] J. Noack, N. Roznyatovskaya, T. Herr, P. Fischer, Angew. Chem. Int. Ed. (2015), http://dx.doi.org/10.1002/anie.201410823. [2] M. Skyllas-Kazacos, M.H. Chakrabarti, S.A. Hajimolana, F.S. Mjalli, M. Saleem,
N. Roznyatovskaya et al. / Journal of Power Sources 302 (2016) 79e83 J. Electrochem. Soc. 158 (2011) R55eR79. €der, I.M. Bayanov, J. Electroanal. Chem. 703 (2013) 29e37. [3] S. Rudolph, U. Schro [4] M.R. Mohamed, H. Ahmad, M.N. Abu Seman, Elektron. Ir. Elektrotech. 19 (2013) 37e42. [5] K. Ngamsai, A. Arpornwichanop, J. Power Sources 282 (2015) 534e543.
83
[6] J. Noack, L. Vorhauser, K. Pinkwart, J. Tübke, ECS Trans. 33 (39) (2011) 3e9. [7] D.N. Buckley, X. Gao, R.P. Lynch, N. Quill, M.J. Leahy, J. Electrochem. Soc. 161 (2014) A524. [8] X. Gao, R.P. Lynch, M.J. Leahy, D.N. Buckley, ECS Trans. 45 (2013) 25e36.
Contents lists available at ScienceDirect
Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour
Short communication
Detection of capacity imbalance in vanadium electrolyte and its electrochemical regeneration for all-vanadium redox-flow batteries Nataliya Roznyatovskaya*, Tatjana Herr, Michael Küttinger, Matthias Fühl, Jens Noack, Karsten Pinkwart, Jens Tübke Fraunhofer Institute for Chemical Technology, Joseph-von-Fraunhofer-Str. 7, Pfinztal 76327, Germany
h i g h l i g h t s UVevis detection of V(III) fraction all-vanadium redox-flow battery electrolyte. Isosbestic point at 600 nm as a reference point for common electrolyte batches. Rebalancing procedure based on pre-charging and remixing.
a r t i c l e i n f o
a b s t r a c t
Article history: Received 19 August 2015 Received in revised form 18 September 2015 Accepted 8 October 2015 Available online xxx
A vanadium electrolyte for redox-flow batteries (VRFB) with different VIII and VIV mole fractions has been studied by UVevis spectroscopy. Spectrophotometric detection enables a rough estimate of the VIV and VIII content, which can be used to detect an electrolyte capacity imbalance, i.e. a deviation in the mole fraction of VIV or VIII away from 50%. The isosbestic point at 600 nm can be used as a reference point in the analysis of common VRFB electrolyte batches. The VRFB electrolyte is observed to have an imbalance after prolonged storage (a couple of years) in a tank under ambient conditions. A regeneration procedure, which involves pre-charging the unbalanced electrolyte and mixing part of it with a portion of initial unbalanced electrolyte, has been tested. The resulting rebalanced electrolyte has been compared with a common electrolyte in a chargeedischarge cell test and is shown to be suitable for cell operation. © 2015 Elsevier B.V. All rights reserved.
Keywords: Vanadium redox-flow battery UVevis absorption Unbalanced state-of-charge Electrolyte
1. Introduction Vanadium salts dissolved in sulfuric acid are used for energy storage in all-vanadium redox-flow batteries (VRFBs) [1,2]. During battery discharge vanadium species undergo redox reactions of VV to VIV in the positive half-cell, and VII to VIII in the negative half-cell. A mixture of VIII and VIV species in a 50:50% mole ratio (referred to as V3.5þ electrolyte) is often used as a commercially-available starting electrolyte for VRFBs, and needs to be charged to form VIII as an anolyte and VIV as a catholyte for further battery operation. During long-term performance the VRFB electrolyte is reported to become unbalanced [3e5]. Imbalance implies an inequality in the amount of oxidized and reduced species in a VRFB, and is usually caused either by vanadium cross-over through the
* Corresponding author. E-mail address: [email protected] (N. Roznyatovskaya). http://dx.doi.org/10.1016/j.jpowsour.2015.10.021 0378-7753/© 2015 Elsevier B.V. All rights reserved.
membrane or by side reactions such as hydrogen evolution in the anolyte. It results in capacity losses, and 100% state-of-charge (SoC) can then be achieved only in one half-cell of the VRFB. To detect the imbalance, the oxidation state and the concentration of vanadium species can be determined and compared. A direct analysis can be performed based on the potentiometric titration of VRFB samples; however, this is time-consuming and unsuitable for continuous electrolyte monitoring or for monitoring online. Indirectly, the SoC can be estimated from its correlation with the electrolyte absorption in the visible range (although this is difficult to use in practice for the catholyte [6]) or from the correlation between the open circuit potential and the vanadium concentration. For the latter technique, a special reference cell is connected to the electrolyte circuit, and the electrical measurement set-up is adapted to it [4,5]. All the above mentioned approaches to detect vanadium electrolyte imbalance concern the analysis of the anolyte and catholyte, i.e. mixtures of VII/VIII and VIV/VV. In this paper we report the imbalance of common V3.5þ electrolyte, which is found to occur during its long-term storage before it is introduced into the battery.
80
N. Roznyatovskaya et al. / Journal of Power Sources 302 (2016) 79e83
As a powerful VRFB system (for example 2 MW/20 MWh) requires several hundred m3 of electrolyte, it is necessary to transport and store batches of electrolyte and to be able to monitor the consistency of their chemical properties over time before they are introduced into the battery. This electrolyte undergoes an alteration of VIII content during storage in a tank over a couple of years, and the ratio of VIII to VIV consequently shifts. As vanadium electrolyte is an important factor in the total costs of the VRFB system, and the published literature on electrolyte rebalancing mostly considers the concentration shifts in the electrolyte during battery operation, we here present an approach for the rapid spectrophotometric identification of imbalance in common V3.5þ electrolyte, and a procedure for the electrolyte pre-treatment to overcome the imbalance. The proposed pre-treatment is based on electrolysis and mixing of electrolyte portions to obtain separate VIII and VIV electrolyte solutions. 2. Experimental 2.1. Chemicals and material Vanadium electrolyte solution (V3.5þ) with a 50:50% mole fraction of VIII and VIV at a total vanadium concentration of 1.55 M, 2 M sulfuric acid and approx. 0.05 M phosphoric acid (GfE Gesellschaft für Elektrometallurgie mbH, Germany) was used as a reference electrolyte, and to prepare samples with various VIII:VIV ratios. Samples of the same V3.5þ electrolyte which had been stored in a tank and exposed to ambient temperature fluctuations over a couple of years were used as an unbalanced electrolyte. The Fumasep FAP-450 membrane (FuMa-Tech GmbH, Germany) was used as the separator. GFA6 graphite felt (SGL Group, Germany) and FU 4369 graphite bipolar plates (Schunk Kohlenstofftechnik GmbH, Germany) were used as electrode current collector materials in a 40 cm2 cell assembly for the chargeedischarge test (for details of cell design, see Ref. [6]). The felts were pre-treated for one hour at 400 C in the oven before the cell assembly, and were compressed by 10% of their thickness in the cell. The electrolyte flow rate was 75e80 ml/min. A 40 cm2 cell of the same design was used for galvanostatic electrolysis to prepare the electrolyte samples with various VIII:VIV ratios and for pre-charging (i.e. rebalancing) of the unbalanced electrolyte. The vanadium concentrations in all the electrolyte samples were determined by potentiometric titration. 2.2. UVevis characterization
out with 0.1 M cerium(IV) sulfate standard solution using Titrator T70 (Mettler Toledo Int. Inc., Germany).
3. Results and discussion 3.1. Visible spectroscopy characterization To model the unbalanced electrolyte samples with different formal charges from 3þ to 4þ but a constant total vanadium concentration, the common V3.5þ electrolyte was galvanostatically charged over a predefined time. These samples were studied using UVevis spectroscopy. The absorbance spectra of these electrolyte solutions with different ratios of VIII to VIV are shown in Fig. 1. The absorption spectrum of neat VIII has two characteristic peaks at 401 and 600 nm. The extinction coefficients of VIII estimated in the concentration range from 0.2 M to 1.6 M are 13.1 M1 cm1 at 401 nm and 7.5 M1 cm1 at 600 nm. The absorption spectrum of neat VIV shows a band positioned at a maximum of 760 nm with a shoulder at 600 nm [7,8]. A linear dependence between the absorption at 760 nm and the VIV concentration has been observed, at least in the range from 0.2 M to 1.6 M. The extinction coefficient of VIV at 760 nm is 19.5 M1 cm1. It can be seen that the spectra of mixtures consisting of VIII and VIV are the convolution of VIII and VIV absorptions. The evolution of spectra with an increase in the VIV fraction results in the appearance of an isosbestic point, which is consistent with the fact that the common electrolyte is a mixture of non-interacting VIII and VIV species and that the analytical concentration of vanadium is constant throughout the series [6]. All three characteristic peaks can be seen for the 50:50% mixture of VIII and VIV. The unbalanced electrolyte displays a lower absorption at the isosbestic point (dashed line, Fig. 1), so the total vanadium concentration in this sample is lower. This is in good agreement with redox titration data: the total vanadium concentration is 1.4 M compared to an initial concentration of 1.55 M. It means that the slow oxidation of VIII by air cannot be the main reason for the shift in the VIII:VIV ratio, and VIII is likely to precipitate out of the electrolyte solution. To be used for quantitative analysis, VIII and VIV bands, which are overlapped on the spectra of the mixtures (curves (2) to (10) in Fig. 1), should be deconvoluted, especially for VIV fractions up to 40%. However, as can be seen in Fig. 2, the ratio of absorbance taken at 760 nm and at 401 nm (Abs760/Abs401) can be used to roughly detect an VIII:VIV ratio shift. The dependence of Abs760/Abs401 in
The absorbance of vanadium electrolyte samples was recorded with a Shimadzu 1650PC UVevis double beam spectrophotometer. These measurements were carried out in matched quartz cuvettes with a 0.2 mm optical path length. 2.3. Electrochemical procedures The charging of electrolyte for rebalancing was carried out galvanostatically at 20 mA cm2 until the redox potential of the anolyte reached approx. 0.5 V vs. Hg/Hg2SO4 electrode. The redox potential of the anolyte was monitored using a glassy carbon electrode incorporated into the electrolyte circuit. The chargeedischarge cell test was performed using a Basytec battery tester (Basytec GmbH, Germany) in galvanostatic mode: the end-ofcharge voltage was set to 1.65 V and the end-of-discharge voltage was 0.8 V. Between the cycles the cell was kept at open circuit potential for five minutes. At each current density (from 25 mA cm2 to 100 mA cm2) five cycles of chargeedischarge were performed. Each chargeedischarge measurement was repeated for reproducibility. The titration of electrolytes samples was carried
Fig. 1. Evolution of absorption spectra in the visible range for vanadium electrolyte as a function of the VIV molar fraction. Total vanadium concentration is 1.55 M for samples (1)e(11), and 1.4 M for the sample (12).
N. Roznyatovskaya et al. / Journal of Power Sources 302 (2016) 79e83
81
the concentration of VIV (65%). This is in good agreement with the titration data, which indicated that the altered electrolyte has the formal vanadium oxidation state þ3.65. Furthermore the spectra of vanadium electrolytes with different VIII:VIV ratios in the range from 0 to 100% can therefore be used for a qualitative estimation of changes in the total concentration of vanadium.
3.2. Electrochemical treatment
Fig. 2. Correlation of the ratio of absorptions at 760 nm and 401 nm, corresponding to characteristic peaks of VIII and VIV, to VIV content.
logarithmic scale on the percentage of VIV is linear over the range from 40% to 90% of VIV, and the Abs760/Abs401 ratio for the unbalanced electrolyte calculated from measured UVevis spectrum (Fig. 1, curve (12)) and plotted in this linear region corresponds to
Generally there are several approaches to restore the VIII:VIV ratio: (i) Adding VIII salt to compensate the loss or decrease in its concentration; this is, however, expensive (VIII is available mainly as vanadium(III) oxide). (ii) Reducing an excess of VIV chemically, i.e. introducing foreign species into the electrolyte (the electrolyte is contaminated by the products of this reaction). (iii) Electrochemical oxidation or reduction of the electrolyte. The electroreduction of an unbalanced electrolyte to V3.5þ can be performed in an anodic halfcell coupled with an oxygen evolution reaction from sulfuric acid solution in the cathodic half-cell. Here special care should be taken in the catalysis of the oxygen evolution reaction. Furthermore it is not always straightforward to set the end point of the electrolysis and achieve the predefined VIII:VIV ratio. In a rebalancing method proposed in Ref. [3] a lack or excess of charge during electrolysis of
Fig. 3. Schematic representation of electrolyte rebalancing procedure. Colors (green, blue, yellow) correspond to the vanadium species in oxidation states þ3, þ4 and þ5. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
82
N. Roznyatovskaya et al. / Journal of Power Sources 302 (2016) 79e83
Table 1 The quantitative composition of vanadium electrolyte samples in the course of pretreatment for rebalancing:a e determined by titration,b e expected, CV e total vanadium concentration. Sample
Anolyte III
Catholyte
V , % V , % CV, M VIII, % VIV, % VV, % CV, M (1) Unbalanced electrolyte 35a (2) After charging 99a 100b (3) After mixing
IV
65a 1a 0b
1.43a 1.46a
35a
1a 0b
65a 75a 70b 99a 100b
25a 30b
1.43a 1.43a 1.44a
the unbalanced electrolyte is compensated by hydrogen or oxygen evolution reactions (electrolysis of water) following the electrochemical conversion of vanadium, to obtain an electrolyte with SoC either 0% or 100%. In this case an addition of water is needed and special oxidation resistant electrodes for cathodic half-cells were used. In this study the pre-treatment of an unbalanced electrolyte comprised three steps: (i) measurement of the VIII:VIV ratio, (ii) initial charging of the electrolyte to generate a 100% mole fraction of VIII in the anodic compartment, whereby a VIV/VV mixture is formed in the cathodic compartment (iii) mixing of the VIV/VV electrolyte with an initial electrolyte sample in a volumetric ratio calculated to ensure that all the VV reacts with VIII to form VIV. This ratio is calculated based on reaction stoichiometry and concentrations of vanadium species in the initial and electrolyzed samples. The procedure is shown in Fig. 3, in which the colors correspond to the vanadium species in different oxidation states presenting in solution. This procedure results in an excessive amount of electrolyte in the form of VIV, which cannot be directly used as an electrolyte, but no additional cell arrangement or special electrodes are needed. The residual amount of VIV catholyte can in principle be used for vanadium recycling. The total vanadium concentration remains the same as in the initial unbalanced electrolyte sample, because this work was focused on rebalancing of the VIII:VIV ratio. Each step of the electrolyte pretreatment was followed by titration to monitor the redox conversion of vanadium species (Table 1). It can be seen that unbalanced anolyte (V3.65þ) is reduced to VIII. The increase in the total vanadium concentration by 2% of its initial value is unexpected. This is probably caused by water diffusion through the membrane, resulting in anolyte volume
change. The catholyte is oxidized to the formal state V4.25þ (75% of VIV and 25% of VV), which is slightly lower than the expected V4.3þ. The reason for this small deviation is unclear. The catholyte is then mixed with initial unbalanced electrolyte in a volume ratio calculated to ensure that all the VV reacts with VIII to form VIV. For the unbalanced electrolyte in this work the ratio was 1:1 v/v. The formal charge of vanadium electrolyte decreased to þ3.99 after the mixing. Both the catholyte and anolyte generated were introduced into the cell for the chargeedischarge test. 3.3. Cell test study As the proton and counter-anion balance is not usually considered during the charging the electrolyte, it is important to know whether the rebalanced electrolyte can be used for cell operation at the same efficiency as the initial V3.5þ electrolyte. Chargeedischarge cell tests were therefore performed with the electrolyte after capacity rebalancing. For comparison the same tests were performed with a cell filled with common (V3.5þ) batch electrolyte taken as a reference. Energy efficiency, coulomb efficiency and discharge power density were evaluated in these tests, and the results are shown in Fig. 4. The cell operated with V3.5þ electrolyte can be discharged at a maximum current density of 50 mA cm2, attaining a discharge power density of 45 mW cm2. The maximum energy efficiency is 76% at 25 mA cm2. From the results shown in Fig. 4 it can be concluded that the cell operated with rebalanced electrolyte demonstrates the same parameters as the cell with standard electrolyte, even though the total vanadium concentration in the rebalanced electrolyte is 1.4 M compared to 1.6 M in the common electrolyte. The difference of 12% in the total vanadium concentration does not, therefore, have a significant influence on the cell performance. 4. Conclusions Vanadium electrolyte consisting of a mixture of VIII and VIV salts in sulfuric acid is found to undergo aging, i.e. an alteration of the VIII mole fraction and total vanadium concentration during storage for a couple of years under ambient conditions, i.e. without maintaining stable temperature and protection against oxygen. This alteration can be detected by spectrophotometry in the visible range. The ratio of adsorption peaks of VIII and VIV in logarithmic scale is a linear function of the VIV mole fraction, at least if VIV exceeds 40%. This ratio can be used to roughly estimate the VIV and VIII content. A rebalancing process for the unbalanced electrolyte has been proposed and examined. The compensation of accumulated VV in the catholyte, after charging the electrolyte to obtain SoC 0%, is carried out by mixing the generated catholyte with initial unbalanced electrolyte. VV then reacts chemically with VIII to form VIV. The mixing protocol can be adjusted for defined VIII:VIV ratios in unbalanced electrolyte. Acknowledgments This work is financially supported by the State of BadenWuerttemberg and the German Federal Ministry of Education and Research (BMBF) in the context of the project “RedoxWind”. The authors are grateful to Mr. Colin Dessornes and Ms. Jessica Kunzendorf for technical assistance in this work. References
Fig. 4. Parameters evaluated from the chargeedischarge tests on the redox-flow battery with common Vþ3.5 electrolyte (solid lines) and capacity rebalanced electrolyte (dashed lines).
[1] J. Noack, N. Roznyatovskaya, T. Herr, P. Fischer, Angew. Chem. Int. Ed. (2015), http://dx.doi.org/10.1002/anie.201410823. [2] M. Skyllas-Kazacos, M.H. Chakrabarti, S.A. Hajimolana, F.S. Mjalli, M. Saleem,
N. Roznyatovskaya et al. / Journal of Power Sources 302 (2016) 79e83 J. Electrochem. Soc. 158 (2011) R55eR79. €der, I.M. Bayanov, J. Electroanal. Chem. 703 (2013) 29e37. [3] S. Rudolph, U. Schro [4] M.R. Mohamed, H. Ahmad, M.N. Abu Seman, Elektron. Ir. Elektrotech. 19 (2013) 37e42. [5] K. Ngamsai, A. Arpornwichanop, J. Power Sources 282 (2015) 534e543.
83
[6] J. Noack, L. Vorhauser, K. Pinkwart, J. Tübke, ECS Trans. 33 (39) (2011) 3e9. [7] D.N. Buckley, X. Gao, R.P. Lynch, N. Quill, M.J. Leahy, J. Electrochem. Soc. 161 (2014) A524. [8] X. Gao, R.P. Lynch, M.J. Leahy, D.N. Buckley, ECS Trans. 45 (2013) 25e36.