Abstract
Vanadium redox flow battery (VRFB) is a promising technology for energy storage because of its independent energy to power ratio and long cycle life. However, VRFB commercialization is still hindered by some technological issues, among which the capacity loss induced by the undesired transport of ions across the ion-exchange membrane. Depending on the nominal operating condition, the choice of a suitable membrane for VRFB application results from the trade-off between low vanadium ions permeability and high proton conductivity. Usually, in order to reduce undesired ion fluxes between the two half-cell electrolytes, the membrane thickness is relatively high, implying increased ohmic losses (i.e., reduced energy efficiency) and system capital cost. In fact, state of the art membranes can represent up to 50% of stack capital cost [1]. Nafion® is widely used due to its high conductivity, but it is not ideally selective towards vanadium ions, leading to the adoption of thicker membrane to limit capacity loss. Alternative cation exchange membranes like SPEEK or SPI are promising because of their reduced permeability, but the low conductivity limits system power density. Instead, anion exchange membranes are still limited by the poor chemical stability and low conductivity [2].
In a recent work [3], the authors demonstrated the proof of concept of an additional selective layer to mitigate vanadium crossover. The selective layer, termed as barrier, is a porous component in which pores size, tortuous path, thickness and composition are designed to improve ion/proton selectivity. The proof of concept was manufactured with reactive spray deposition technology (RSDT), which is a flame-based synthesis process unique to Dr. Radenka Maric's research group. For the fabrication of the proof of concept, carbon-rich particles ∼4-10 nm in diameter were formed in the RSDT flame and were deposited directly onto Nafion® 212 (50 μm thick) simultaneously with a mixture of 1100EW Nafion® and Vulcan® XC-72R (∼40 nm diameter) that was sprayed from air-assisted secondary nozzles. The presence of the barrier layer significantly reduced battery self-discharge, as reported in [3].
In this work, different compositions and morphologies of the barrier layers were analysed in order to improve ion/proton selectivity. In particular, the influence of ionomer to carbon ratio (I/C), the amount of carbon-rich particles and the introduction of silica were investigated. Moreover, the effect of Nafion® 211 (25 μm thick) as a support for barrier deposition was also evaluated. The barrier layers were characterized in a 25 cm2 cell [4] equipped with reference electrodes at both positive and negative electrode in order to monitor the corresponding electrolyte potential and get an insight into battery state of charge (SoC). In addition to electrochemical testing, the structure of the barrier layer was characterized using TEM and SEM.
The most suitable I/C ratio was found to be included between 1 and 2, while the presence of carbon-rich particles significantly contributed to crossover reduction, with a minor impact on proton conductivity. Also the introduction of silica nanoparticles from primary nozzle was effective for vanadium ion selectivity, in particular when the amount of carbon from the secondary nozzle is reduced. In fact, the most promising barrier layer resulted the one composed by only silica and ionomer. This layer was also deposited on Nafion® 211 (Figure 1), exhibiting an excellent trade-off between ion selectivity and proton conductivity. This barrier was proved to be stable over 1,000 cycles, presenting a stable coulombic efficiency of 99.5%, with an average capacity decay of 0.08%/cycle at 100 mA cm-2.
Figure 1 – SEM image of only Silica barrier deposited on Nafion® 211.
References:
[1] C. Minke et al., Journal of Power Sources 376 (2018) 66-81.
[2] Y. Shi et al., Applied Energy 238 (2019) 202-224.
[3] M. Cecchetti et al., Journal of the Electrochemical Society 167 (2020) 130535.
[4] M. Cecchetti et al., Journal of Power Sources 400 (2018) 218-224.
Figure 1