Electrode Treatments for Redox Flow Batteries: Translating Our Understanding from Vanadium to Aqueous-Organic

Abstract

Redox flow batteries (RFBs) are a promising technology for long-duration energy storage; but they suffer from inefficiencies in part due to the overvoltages at the electrode surface. In this work, more than 70 electrode treatments are reviewed that are previously shown to reduce the overvoltages and improve performance for vanadium RFBs (VRFBs), the most commercialized RFB technology. However, identifying treatments that improve performance the most and whether they are industrially implementable is challenging. This study attempts to address this challenge by comparing treatments under similar operating conditions and accounting for the treatment process complexity. The different treatments are compared at laboratory and industrial scale based on criteria for VRFB performance, treatment stability, economic feasibility, and ease of industrial implementation. Thermal, plasma, electrochemical oxidation, CO₂ treatments, as well as Bi, Ag, and Cu catalysts loaded on electrodes are identified as the most promising for adoption in large scale VRFBs. The similarity in electrode treatments for aqueous-organic RFBs (AORFBs) and VRFBs is also identified. The need of standardization in RFBs testing along with fundamental studies to understand charge transfer reactions in redox active species used in RFBs moving forward is emphasized.

Keywords: aqueous-organic redox flow batteries; carbon felts; electrocatalysts; electrode treatments; flow batteries; quinones; vanadium.

Figure 1

Schematic of a Redox Flow Battery. Carbon felts are commonly used as the negative and positive electrodes. The schematic is not drawn to scale.

Figure 2

Illustrative plots of a) Cell voltage versus current density for two different electrode treatments (Treatment 1 and 2) in both charging and discharging mode. The different overvoltage contributions for Treatment 1 are shown. b) Comparing energy efficiency for the two treatments in part (a) at a particular current density, assuming coulombic efficiency remains unchanged.

Figure 3

Proposed charge transfer mechanisms for redox couples in vanadium redox flow batteries during charging in sulfuric acid on carbon electrodes considering solvation of vanadium ions, (a) V³⁺/V²⁺ and b) VO²⁺/VO₂ ⁺. The V³⁺/V²⁺ mechanism is based on the findings of references^{[ 17} , ^{26 ]} and VO²⁺/VO₂ ⁺ is based on references.^{[ 53} , ⁵⁶ , ⁵⁷ , ⁵⁸ , ^{59 ]} The starting complex structures are based on findings of references.^{[ 25} , ⁶⁰ , ⁶¹ , ⁶² , ⁶³ , ⁶⁴ , ⁶⁵ , ⁶⁶ , ^{67 ]} The exact charge of the adsorbed complex is unknown and so a best approximation is given, consistent with charge balance.

Figure 4

Energy efficiency versus current density (i) relationships of VRFB with treatment ID Thermal‐2 with treated (T) and untreated CFs (UT) along with scaled‐up VRFB (200 kW/ 400 kWh) with untreated CFs (UT/S). Trendlines fitting the data are shown by dotted lines. R² is the coefficient of determination. The current densities and energy efficiencies of VRFB with treatment ID Thermal‐2 with T and UT CFs is extracted from reference^{[ 69 ]} and scaled‐up VRFB (200 kW/ 400 kWh) with UT/S CFs is extracted from reference.^{[ 149 ]}

Figure 5

a) Relative change in EE of VRFB using treated (T) CFs with treatment ID Thermal‐2 with respect to untreated (UT) CFs at operating current densities of 50, 75, and 100 mA cm⁻². The dotted line shows the nominal 5% change required to satisfy the criteria. b) Absolute energy efficiency (EE) of VRFB with CFs with treatment ID Thermal‐2 and scaled‐up VRFB (200 kW/ 400 kWh) with untreated CFs (UT/S) at different operating current densities of 50, 75, and 100 mA cm⁻². The raw current densities and energy efficiencies used to obtain performance parameters at selected current densities of 50, 75, and 100 mA cm⁻² of VRFB with treatment ID Thermal‐2 with T and UT CFs is taken from reference^{[ 69 ]} and scaled‐up VRFB (200 kW/ 400 kWh) with UT/S CFs is taken from reference.^{[ 149 ]}

Figure 6

a) Flowchart showing the steps followed to evaluate the affordable capital cost (ACC). Here N_{Cells, T} is the total number of cells needed to deliver the rated power for VRFB with treated CFs and A_electrode is the area of electrode used on each side of the battery. b) Example plot showing the capital expenditure per unit area of electrode for VRFBs with UT/S, UT, and T CFs used to evaluate ACC_UT/S and ACC_UT at a selected energy efficiency.

Figure 7

Affordable capital cost for VRFB with ID Thermal‐2 at energy efficiencies of 67.3, 73, and 78%. The dotted line indicates the minimum value of ACC required for a treatment to satisfy the criteria. The ACC for VRFB with ID Thermal‐2 at selected energy efficiencies of 67.3, 73, and 78% utilizes the current densities and energy efficiencies relationships developed using performance data obtained for VRFB with treatment ID Thermal‐2 with T and UT CFs from reference^{[ 69 ]} and scaled‐up VRFB (200 kW/ 400 kWh) with UT/S CFs from reference.^{[ 149 ]}

Figure 8

a) Carbon felt treatments that satisfy (GREEN) or do not satisfy (RED) the performance, stability, and economic feasibility criteria. Inadequate information for a treatment is shown by GRAY color. See Table S1 (Supporting Information) for ID information and text for criteria and b) Carbon felts treatments sorted in various categories based on whether they satisfy the criteria for performance, stability, and economic feasibility. The allowed color code combinations for a treatment to belong to a particular category is shown by colored squares on the right of each category.

Figure 9

a) Metal and metal oxide electrocatalysts that satisfy (GREEN) or do not satisfy (RED) the performance, stability, and economic feasibility criteria. Inadequate information is shown by GRAY color. See Table S2 (Supporting Information) for ID information and text for criteria and b) Metal and metal oxide electrocatalysts sorted in various categories based on whether they satisfy the criteria for performance, stability, and economic feasibility. The allowed color code combinations for a treatment to belong to a particular category is shown by colored squares on the right of each category.

Figure 10

Carbon felt treatments and metal and metal oxide electrocatalysts grouped based on their level of complexity. The treatment complexity is the number of steps in the process and process units required to implement at the industrial scale. The total number and details of steps and process units needed for each treatment are available in Tables S3 and S4 (Supporting Information).

Figure 11

Different stages of making carbon fibers from PAN precursors. Reproduced with permission.^{[ 169 ]} Copyright 2017, The American Society of Mechanical Engineers.

Figure 12

Cost associated with each stage for production of carbon fibers, assuming 1500 ton year⁻¹ production capacity. The area of the squares or rectangles and the text size for each of the steps is proportional to their relative costs. The details regarding assumptions are available in references.^{[ 170} , ¹⁷² , ¹⁷³ , ^{174 ]}

Figure 13

Preparation of carbon felts from PAN precursors. The ratio of cost of textile and carbonization process relative to cost of precursor is shown in green and red colored circles, respectively. Here f denotes the cost of the PAN precursor material. Reproduced with permission.^{[ 175 ]} Copyright 2017, Elsevier.

Figure 14

a) Metal loading and b) Cost of metal salt used for various metal electrocatalysts in RFBs, and c) Total capital cost associated with loading these metal electrocatalysts per unit area of carbon felts. The metal loading for Bi, Cu, and Ag is obtained from references,^{[ 176 ]},^{[ 121 ]} and^{[ 120 ]} respectively. The costs of the metal salts have been obtained from vendors that can supply these in bulk for meeting industrial requirements.

Figure 15

Total capital cost associated with treatments for VRFBs that improve performance or stability, is economically feasible, and are easy to implement.

Figure 16

Classes of organic molecules used in aqueous‐organic redox flow batteries based on their structure. A representative molecule for each class is shown, DMBQ for quinones, K₄Fe(CN)₆ for organometallic coordination complexes, TMAP‐TEMPO for nitroxide radical derivatives, BTMAP‐Vi for viologens, and MB for all other aromatic heterocycles. Here DMBQ = 2,6‐dimethoxybenzoquinone, K₄[Fe(CN)₆] = Potassium hexacyanoferrate, TMAP‐TEMPO = 4‐[3‐(trimethylammonio)propoxy]−2,2,6,6‐tetramethylpiperidin‐1‐oxyl, BTMAP‐Vi = bis(3‐trimethylammonio)propyl viologen tetrachloride, and MB = Methylene Blue.