Techno-Economic Materials-Selection Criteria for Non-Aqueous Flow-Battery Chemistries

Cost projections have predicted that non-aqueous flow-batteries (NRFBs) having high molar-equivalents could enable electrochemical energy-storage on the Grid [1]. One approach is to use high-solubility redox-active molecules in solution.  Key material parameters in the design of redox-active molecules are molecular weight and the potential at which redox occurs.  Here, we use a previously developed techno-economic model [1] to determine targets for the properties of redox-active molecules that are necessary to satisfy system-price targets.  This analysis links system-level goals for cost directly to material properties and provides a means to screen materials (1) developed by chemical synthesis and (2) theorized through first-principles modeling.

The present techno-economic model distributes NRFB system-price (in $/kWh) among three parts: reactor, electrolyte, and additional costs.  Additional costs are insensitive to the particular choice of redox-active species and have been estimated at $63/kWh [1].  The equilibrium potential of the full-cell affects the reactor cost, and, therefore, the cost of a single chemistry cannot be evaluated without introducing an appropriate counter-electrode.  To analyze candidate anolyte chemistries we consider a hypothetical catholyte (with which the anolyte pairs) having identical molecular weight and solubility as the chemistry of interest, but having an equilibrium potential of 4 V vs. Li⁰/Li⁺.  We analyze catholyte chemistries in an analogous way having a hypothetical anolyte with 1 V vs. Li⁺/Li⁰ potential.  Molecular weight enters into the system price through the equivalent weight, defined as the molecular weight per electron transferred.  Reaction potential measurements and the number of redox events (which affects equivalent weight) can be determined rapidly from cyclic voltammetry.  In contrast, the measurement of active-species solubility (which is known to reduce system price [1]) requires preparation of multiple samples or a temperature sweep.  Therefore, it is desirable to know the active-species solubility level that is required to produce a certain system-price – in this case $120/kWh.  

Because of the availability of potential and molecular-weight measurements and the relative scarcity of solubility measurements, we have developed a target-solubility map (Fig. 1) to define the design-space for redox-active molecules in NRFBs.  In this space a particular chemistry is identified by its reaction potential (y-axis in V vs. Li⁺/Li⁰) and equivalent weight (x-axis in g/mol-e^-).  Here, we do not emphasize specific materials of interest, but, instead, the properties of materials (reaction potential, equivalent weight, and solubility) that are necessary to reach a system price of $120/kWh.  The red contour lines represent just that when a low-potential counterelectrode (1V vs. Li⁺/Li⁰) is assumed with a particular active-species solubility level (in kg of active-species per kg of electrolyte, i.e., kg:kg).  Blue contour lines represent analogous cases when a high-potential counterelectrode (4V vs. Li⁺/Li⁰) is assumed.  Examination of the feasible space defined by these materials-selection criteria reveals useful insights into molecular design, including that (1) high equivalent-weight molecules can be tolerated if reaction potentials are extreme (i.e., <1V vs. Li⁺/Li⁰ or >4V vs. Li⁺/Li⁰) and (2) a “no man’s land” exists for which ultralight molecules cannot meet the system-price target even with infinite solubility when these molecules produce small cell-potentials.  In addition, we examine these characteristics for materials that exhibit multiple redox events, in which case these criteria provide economic guidelines for which redox events to use when operating a cell.  Furthermore, we examine the sensitivity of the molecular design-space to rate-capability parameters, including electrolyte viscosity and ionic conductivity.

1. R.M. Darling, K.G. Gallagher, J.A. Kowalski, S. Ha, and F.R. Brushett, Energy Environ. Sci., 7, 3459 (2014).