Non-Aqueous Single-Metal Redox Flow Batteries.

Abstract

Redox flow batteries (RFBs) are being developed for large-scale energy storage and load-leveling systems for solar or wind power. Aqueous chemistries are used for current commercial RFBs, in which energy and power density are limited by the stability range of water. Non-aqueous solvents offer stability windows up to four times greater than those for aqueous solvents. The goal of my research was to examine all components of a non-aqueous-vanadium-single-metal RFB and determine their effects on key performance characteristics.

First, relationships between the structure, composition, and function of acetylacetonate metal complex based electrolytes were examined in an effort to determine strategies for their further development and provide initial guidelines for their use. Vanadium, chromium, and manganese acetylacetonate complexes had maximum energy densities of 18, 18, and 9 Wh/L respectively with reversible electrochemistry for V and Mn; therefore vanadium was selected for more extensive testing. Substitution of the ligands demonstrated the ability to change solubility by two orders of magnitude. Results from examination of several solvent/supporting electrolyte combinations indicated that solvents with low solvent molar volumes and high polarities possessed desirable properties (acetonitrile is optimal).

Effects of the cell components (membrane resistance and electrode kinetics) on the cell performance were also examined. The lowest resistance membranes, Selemion DSV or Neosepta AHA, were chosen to reduce energy losses. The kinetics of the desired reaction on gold, platinum, and glassy carbon electrodes showed minimal kinetic limitations suggesting outer-shell-electron-transfer reactions occur.

Finally, the stability was examined. When exposed to water or oxygen, the V(II)/V(III) redox couple becomes irreversible and vanadyl acetylacetonate is formed. Even in the absence of oxygen or water impurities, the capacity of the RFB fades dramatically. This fade could be a consequence of precipitation stemming from a reaction between the charged active species and the acetonitrile solvent. Overpotentials on the electrode and membrane increased with cycling – likely due to precipitation and mechanical degradation and could contribute to capacity fade (based on results from scanning electron microscopy).

Overall I found that the non-aqueous all-vanadium RFB could be a promising candidate for future batteries after stability of the cell components is addressed.