There is considerable interest in flow batteries for storing energy from non-dispatchable power sources such as solar, wind and ocean energy and for other large and medium scale energy storage applications.  Vanadium flow batteries (VFB), also known as vanadium redox flow batteries (VRFB or VRB), are particularly attractive because, in addition to having long cycle life, they are essentially immune to cross-contamination problems due to mass transfer across the membrane that can limit the service life of the electrolyte in other systems.

 Accurate monitoring of state of charge (SoC) is intrinsically important for the reliability of energy storage systems, particularly large systems in critical applications.  Furthermore, independent monitoring of the SoC of both electrolytes is important for effective operation of flow battery technology.  For example, in a VFB, transfer of vanadium ions across the membrane and side reactions such as hydrogen formation at the negative electrode can cause the battery to become unbalanced (e.g. more V^V on the positive side than V^IIon the negative), and balancing by overcharging is not practical because it can lead to significant degradation of the carbon at the positive electrode.

The principal active species in a VFB are vanadyl (VO²⁺) and pervanadyl (VO₂⁺) ions (i.e. V^IV and V^V) in the positive electrolyte (catholyte) and V³⁺ and V²⁺ in the negative electrolyte (anolyte), typically in H₂SO₄ solutions.   These solutions are highly colored: the vanadium species all have strong absorbance spectra in the visible region. Thus, as we have previously suggested,^1,2 ultraviolet-visible (UV-Vis) spectroscopy offers an attractive method of independently measuring the SoC of both electrolytes.  Spectroscopic monitoring of SoC is independent of electrochemistry and offers the possibility of performing in-situ analysis.  Because the absorbance of V^II-V^III mixtures is a linear combination of that of the constituents, it is straightforward to implement UV-Vis spectroscopic monitoring of the SoC of the anolyte. However, at the concentrations typically used in a VFB, the absorbance of V^IV-V^V electrolytes is high and is a very non-linear function of the mole fraction of V^IV and of overall vanadium concentration.^1-5  Because of this, it was previously suggested in the literature that UV-Vis spectroscopic monitoring of VFB catholytes is not feasible.

We have shown^4,5 that the non-linear absorbance behavior of the catholyte can be quantitatively explained by the formation of a strongly absorbing 1:1 mixed-valence complex, V₂O₃³⁺, in equilibrium with VO²⁺ and VO₂⁺.  A model of the spectra based on this equilibrium, using an excess absorbance parameter p to quantify the effect of V₂O₃³⁺ formation together with the extinction coefficients of VO²⁺ and VO₂⁺, shows excellent quantitative agreement with experiment and comprehensively explains the spectroscopic behaviour.  In this paper, we present results for the UV-Vis spectroscopy of VFB catholytes over a range of concentrations and extend our model to analyse electrolytes with different concentrations of total vanadium.

We measured and analysed the ultraviolet-visible spectra of catholytes for vanadium flow batteries (VFBs) for a range of V^IV:V^V ratios and vanadium concentrations.  The results showed that p varies weakly with the vanadium concentration C and we quantified this variation relative to a reference concentration C_r  in terms of a single parameter M  that is independent of the choice of C_r.  We generated standard spectra of p (at a reference concentration of 1 mol dm^-3) and of the extinction coefficients of  VO²⁺ and VO₂⁺.  The spectrum of any catholyte may be simulated using the measured value of M in a governing equation which expresses the absorbance, at any wavelength λ, of catholyte with any V^IV mole fraction f and any total vanadium concentration C in terms of these standard spectral parameters.  The precision of the simulation was verified against experimental spectra of an independent set of V^IV-V^Vmixtures.

We constructed calibration curves for SoC using the governing equation and the standard spectral parameters.  Values of SoC determined from the measured absorbance of independent V^IV-V^V mixtures using these calibration curves were in good agreement with the corresponding values determined from the concentrations of the parent solutions.  Based on these results, we extended our experiments to in-situmonitoring of SoC by UV-Vis spectroscopy.   We will discuss our results with particular emphasis on the behavior of the positive electrolyte. 

1. X. Gao, et al., ECS Trans., 45, 25 (2013).

2. D.N. Buckley, et al, European Patent EP 13195315 (2013).

3. P. Blanc, et al, Inorg. Chem., 21, 2923 (1982).

4. D.N. Buckley, et al, J. Electrochem. Soc., 161, A524 (2014).

5. C. Petchsingh, et al, J. Electrochem. Soc. 163, A5068 (2016)