Exploring Rate Capabilities of Symmetric Nonaqueous Redox Flow Batteries

Abstract

After decades of research, widespread implementation of redox flow batteries (RFBs) for grid electricity storage has only recently begun to appear feasible. With MWh-level installations in the U.S. and GWh-level installations planned in China [1], the aqueous vanadium RFB chemistry appears to be the market leader. These projects offer levelized cost of capacity lower than the large Li-ion storage systems offered by suppliers like Tesla. A recent US study, which used the state of Michigan’s electrical grid as a test case, found that RFBs reduced emissions per kWh in all scenarios, but also found that RFBs were only economically preferable to wind curtailment at curtailment levels above 66% [2]. So there remains a need to find alternative, less costly energy storage strategies to better compete with other carbon-intensity improvement strategies that are available to utilities. Symmetric RFBs offer a step-change reduction in reactor cost because their use of identical solutions in the discharged posolyte and negolyte mitigates the need for expensive ion-exchange-membrane separators. Many symmetric RFB chemistries exist: redox-active organometallic chemistries such as V(acac)3 ­[3], or redox-active organic chemistries such as polycyclic aromatic hydrocarbons [4]. These RFB chemistries operate at voltages that split water, requiring nonaqueous environments. Nonaqueous RFBs are often rejected out of hand, based on their expected inability to provide reasonable power. Such an expectation is based on a qualitative argument that nonaqueous solutions tend to have low conductivites, and consequently cannot support high-current operation. However, ion-exchange membranes in aqueous solutions also tend to have similarly low conductivities. Being freed from the need for costly, resistive ion-exchange membranes, symmetric chemistries can employ inexpensive porous polyolefin separators, which we will show can permit rate performance comparable to aqueous vanadium RFBs. Results were collected from a bespoke system involving a 6.25 cm2 flow-through reactor and 100 mL active-liquid storage vessels: the positive and negative flow-through regions contained carbon felt at 6% compression to fit the 3 mm depth. Flow cell designs of these nonaqueous systems are non-trivial since many of the materials used in these nonaqueous chemistries are incompatible with materials used for aqueous systems. We will discuss some design criteria and lessons learned in the fabrication of these nonaqueous flow cells. Figure 1 shows 20 cycles of a symmetric V(acac)3 cell operating at 20 mAcm–2 in a cell with a porous polyolefin separator. The voltage cutoff was not triggered during the charging procedure, allowing the charging cycles to achieve the full programmed capacity. The coulombic efficiency, voltage efficiency, and energy efficiency quickly stabilize to 83%, 87%, and 72%, respectively. Informed by similar data across a range of charge/discharge currents and flow rates, we will report more generally on the performance of nonaqueous symmetric RFB chemistries. [1] Unienergy Technologies. www.uetechnologies.com. Accessed Dec. 2016. [2] M. Arbabzadeh, et al. Appl. Energy 2015, 146, 397. [3] Q. Liu et al. Electrochem. Comm. 11 (2009): 2312. [4] J. Saraidaridis et al. ECS Meeting on Electrochemical Energy and Storage s. (2015): 685. Figure 1: 0.1 M V(acac)3/0.5 M TEABF4 in acetonitrile, 20 mAcm-2, polypropylene separator, 50% SOC charging cycles after 1st charge to 75%, 50% SOC 1st discharge and to 1 V cutoff on subsequent cycles, Argon atmosphere with <1ppm O2, H2O Figure 1