Thermal Hydrogen Treatment of Electrodes for Performance and Stability Enhancement of Non-Aqueous Redox Flow Batteries

Abstract

Vanadium acetylacetonate [(V(acac)3] is being used as model redox species for the study of non-aqueous redox flow battery (NAqRFB). The redox species present reversible redox reactions for both V+2/V+3 (i.e. the negolyte ca. -1.75 V vs Ag/Ag+) and V+3/V+4 (i.e. posolyte ca. 0.45 V vs Ag/Ag+). The same discharge state V+3 of this electrolyte can be used on both sides of the half cell. The latter and the fact it dissolves easily in an organic solvent, besides its commercial availability at a fair price makes it a promising candidate for future commercial versions of non-aqueous redox flow batteries.Nonetheless, merits are usually accompanied by some demerits. Vanadium has four valency states starting from V+2 to V+5; for the V-NAqRFB only three oxidations states are desirable V+2, V3+, and V4+ thus the formation of V5+ is unwanted. To avoid this oxidation state, typically the battery open circuit potential is kept below 2.27 V. Another downside of this electrolyte refers to its sensitivity towards a moisture/oxygen-enriched environment. The presence of moisture/oxygen in the environment triggers a non-reversible reaction which leads to the formation of VO(acac)2 at the positive electrode. VO(acac)2 irreversible formation causes rapid battery capacity loss due to its parasitic and increased concentration on the electrolyte. Therefore, the electrolyte must be kept in an inert atmosphere such as dry nitrogen. Nonetheless, it was observed that even by maintaining a strictly controlled saturation of the electrolyte with inert gas, the stability of the batteries was still rather negligible. Components such as the electrodes (carbon felts, CF) could also represent a major source for the oxygen-enriched surface. Many oxidation pre-treatments such as chemical exfoliation (H2SO4) and thermal oxidation under airflow at 450 °C are commonly performed on CFs to increase their wettability and performance in aqueous all vanadium redox flow cells. Nonetheless, these kind of pre-treatments might not be suitable for the system here in the study due to the beforehand mentioned reasons. The latter served as the main motivation for this study.Initially, cyclic voltammetry (CV) was performed using a three-electrode configuration cell and a glassy carbon as the working electrode. The addition of few DI water drop s (0.005% or 50 ppm) and the electrolyte saturation with O2 caused the immediate disappearance of redox species peaks (V2+/V3+ redox couple) and the appearance of additional peaks which indicated a strong influence for the reaction shift towards the formation of parasitic VO(acac)2. V+3/V+4 redox couple was also degraded significantly after 0.05% (500 ppm). Thus, keeping a dry and oxygen-free atmosphere is of utmost importance for maintaining a stable battery operation. Interestingly, it was found that the reversibility of species on top of the WE depends on the exposure time to oxygen. The effect is negligible up to 20 minutes of oxygen saturation but extremely critical over 4 hours as even upon saturating the electrolyte with an inert gas again, the adsorption of the active species was less effective and non-reversible (much lower anodic current peaks and much wider apart oxidation/reduction peaks).The carbon felt electrodes used for the aqueous vanadium redox flow battery, even with no treatment have already oxygen functional groups which serve as the reaction sites for the vanadium redox reaction in the aqueous electrolyte. Hydrogen is known as a powerful reducing agent and so might be employed to remove these oxygen functionalities. Therefore, the impact of different pre-treatments on the carbon felts morphology and performance for the V-NAqRFB was assessed. Mainly, the effect of solely adding a pre-oxygen heat treatment (referred to as PO) to a pristine carbon felt from SGL (4.6GFD), a hydrogen treatment (PH), and a combination of oxygen followed by a hydrogen treatment (POH) to completely remove all previously formed heteroatoms, were tested, each for 10 hours at 400 °C. For comparison purposes, a pristine electrode (P) was also used. Then upon, the electrochemical surface area of each CF was pre-screened by leading CV measurements and the hydrogen treatment followed by an oxygen heat treatment (POH) appeared to be the best solution to increase the number of active sites on the CF structure. Morphological characterization was performed on all electrodes using SEM and FTIR. Later, these CFs were used in NAqRFB-cell for battery cycling, impedance spectroscopy, and polarization measurements. The co-relation of the results highlighted that POH is performing better in the V(acac)3 based non-aqueous electrolyte. Besides, solely conducting a pre-oxygen heat treatment on the CFs condemns the performance of the battery, as expected. The results are shown in the attached figure. Figure 1