Toward a Low-Cost Alkaline Zinc-Iron Flow Battery with a Polybenzimidazole Custom Membrane for Stationary Energy Storage.

Alkaline zinc-iron flow battery is a promising technology for electrochemical energy storage. In this study, we present a high-performance alkaline zinc-iron flow battery in combination with a self-made, low-cost membrane with high mechanical stability and a 3D porous carbon felt electrode. The membrane could provide high hydroxyl ion conductivity while resisting zinc dendrites well owing to its high mechanical stability. The 3D porous carbon felt could serve as a guidance for the zinc stripping/plating, which can effectively suppress zinc dendrite/accumulation as well. Thus this battery demonstrates a coulombic efficiency of 99.5% and an energy efficiency of 82.8% at 160 mA cm?2, which is the highest value among recently reported flow battery systems. The battery can stably run for more than 500 cycles, showing very good stability. Most importantly, the practicability of this battery is confirmed by assembling a kilowatt cell stack with capital cost under $90/kWh.

Introduction

Flow batteries are of tremendous importance for their application in increasing the quality and stability of the electricity generated by renewable energies like wind or solar power (Yang et al., 2011, Dunn et al., 2011). However, research into flow battery systems based on zinc/bromine, iron/chromium, and all-vanadium redox pairs, to name but a few, has encountered numerous problems, such as the corrosion of bromine, poor kinetics of Cr2+/Cr3+ redox pair, relatively high cost, and low energy density of all-vanadium redox pairs, although these battery systems are currently at the demonstration stage (Yuan et al., 2016b, Park et al., 2016). These barriers have, on the one hand, hindered their further wide scale deployment, and on the other hand, accelerated research efforts into new flow battery chemistries (or the next-generation flow batteries, aqueous or non-aqueous redox flow batteries) (Perry and Weber, 2015, Park et al., 2016). Among the reported new systems, non-aqueous redox flow battery systems, having the features of wide electrochemical window, high energy density, inexpensive redox active materials, etc., are currently at the proof-of-concept stage. However, the low concentration and poor ion conductivity of organic-based electrolytes are the most critical issues to overcome (Park et al., 2016). Although aqueous flow battery systems, like TEMPO (2,2,6,6-tetramethylpiperidine-1-oxyl)-based flow battery and quinone-based flow battery, have been successfully demonstrated at the laboratory scale (the electrode area is normally less than 10 cm2), their relatively low performance at high current density (normally less than 100 mA cm−2, when the energy efficiency [EE] was above 80%) (Liu et al., 2016, Wang et al., 2016, Janoschka et al., 2015, Janoschka et al., 2016, Winsberg et al., 2016, Hu et al., 2017) limits the quick response for energy conversion and increases the integration cost. In addition, some of the systems still have low open cell voltage (OCV) or low electrolyte concentration. The low working current density together with the low OCV will result in low power density of a flow battery, further leading to an increased stack size and an overall increased capital cost of a flow battery system. Currently, only a few membrane materials (such as perfluorinated ion-exchange membranes because of their high stability in critical medium, e.g., strongly acidic or alkaline conditions and highly oxidative medium) have been considered (Gong et al., 2015, Gong et al., 2016, Wang et al., 2011, Li et al., 2015, Li et al., 2016, Lin et al., 2015, Lin et al., 2016, Orita et al., 2016), which will definitely further result in cost issue for battery stacks (Chu et al., 2017). Besides, upscaling for practical application of these newly developed aqueous flow battery systems have been rarely reported, which may induce limitations on the energy storage system development to some extent.

The alkaline zinc ferricyanide flow battery owns the features of low cost and high voltage together with two-electron-redox properties, resulting in high capacity (McBreen, 1984, Adams et al., 1979, Adams, 1979). The alkaline zinc ferricyanide flow battery was first reported by G. B. Adams et al. in 1981; however, further work on this type of flow battery has been broken off, owing to its very poor cycle life and the relatively low operating current density (35 mA cm−2) (McBreen, 1984). The poor cycle life is mainly due to the zinc dendrite under alkaline medium, where a cadmium-plated (Zn- or Cu-plated) iron substrate was employed for zinc stripping/plating, while the low operating current density could be due to the high resistance of a cation-conducting membrane and severe zinc dendrite derived from the metal electrode.

Here we present a long cycle life alkaline zinc-iron flow battery with a very high performance. The battery employs Zn(OH)42−/Zn and Fe(CN)63−/Fe(CN)64− as the negative and positive redox couples, respectively, while a self-made, cost-effective polybenzimidazole (PBI) membrane and a 3D carbon felt electrode were combined. The PBI membrane carrying heterocyclic rings can guarantee fast transportation of hydroxyl ions after doping with a base solution (Li et al., 2003, Yuan et al., 2016a). Most importantly, the PBI membrane with ultra-high mechanical stability can resist the zinc dendrite very well, which ensures the cycling stability of the alkaline zinc-iron flow battery. In addition, a 3D porous carbon felt with high porosity and surface area, which serves as guidance for the zinc stripping/plating and suppresses zinc dendrite/accumulation effectively, provides the battery with excellent cycling stability and rate performance. Moreover, the concentration of Fe(CN)63−/Fe(CN)64− redox couple can reach 1 mol L−1 by optimizing the composition of electrolyte, which is much higher than the reported concentration of this redox couple (0.4 mol L−1) (Orita et al., 2016, Lin et al., 2015, Lin et al., 2016, Selverston et al., 2016, Selverston et al., 2017). The high concentration of active materials thus can afford the battery with high energy density. As a result, the proposed zinc-iron flow battery demonstrated an EE of 82.78% even at a high current density of 160 mA cm−2. A charge/discharge experiment of 500 cycles further confirmed the excellent stability of this system.

Results and Discussion

Electrochemical Performance of the Alkaline Zinc-Iron Flow Battery Using a PBI Membrane and a 3D Porous Carbon Felt Electrode

Figure 1A showed the principle and structure of the reported alkaline zinc-iron flow battery, where Zn(OH)42−/Zn pair served as the negative active material and Fe(CN)63−/Fe(CN)64− pair was employed as a positive redox couple. For batteries involving stripping/plating process, especially for Li- and zinc-based batteries, a membrane with high mechanical stability is always in high demand (Figure 1C), since it can effectively suppress dendritic growth during metal plating (Li or zinc), whose safety issues upon internal short circuit could be mitigated. A PBI membrane (Figure S1), which is widely used in fuel cells because of its very high chemical and mechanical stability (the elasticity modulus of the PBI was higher than 2.9 GPa, Figure S2), along with very high thermal stability, was used as the separator. Carrying heterocyclic rings, a PBI membrane can guarantee the fast transportation of hydroxyl ions after doping with a base solution (Figure 1D). For comparison, a Nafion (Nafion 115, whose elasticity modulus was about 94 MPa, Figure S2) membrane, which was widely used in aqueous flow battery systems, even in the alkaline medium, was used as a reference.

Figure 1

The Utilization of an Alkaline Zinc-Iron Flow Battery Using a Self-made, Low-Cost Non-fluorinated Ion-Exchange Membrane

(A) Schematic of the alkaline zinc-iron flow battery.

(B) Cyclic voltammetry of negative and positive redox pairs in sodium hydroxide solution. The negative electrolyte contains 0.1 mol L−1 Zn(OH)42− and 3 mol L−1 sodium hydroxide solution; the positive electrolyte contains 0.2 mol L−1 K3Fe(CN)6, 0.2 mol L−1 K4Fe(CN)6, and 3 mol L−1 sodium hydroxide solution. The working electrode is graphite plate, and the scan rate is 40 mV s−1.

(C) Schematic of introjecting a PBI membrane with ultra-high mechanical stability and a 3D porous carbon felt for an alkaline zinc-iron flow battery.

(D) The transportation mechanism of OH− in a PBI membrane.

See also Figures S1–S4, S8–S16, S18, S20, and S21.

The cyclic voltammetry (CV) of the two redox pairs on a graphite plate working electrode at a scan rate of 40 mV s−1 (Figure 1B) predicts an equilibrium formal potential of 1.74 V, which corresponds to the redox reactions Fe(CN)63−/Fe(CN)64− at the positive side and Zn(OH)42−/Zn at the negative side. CV curves at different scan rates ranging from 10 mV s−1 to 60 mV s−1 reveal that Fe(CN)63−/Fe(CN)64− redox reaction is a standard diffusion-controlled process (Figure S3A). Based on the Randles-Sevcik equation, the diffusion coefficients of Fe(CN)63− and Fe(CN)64− in the alkaline electrolyte were found to be 9.03×10−6 cm2 s−1 and 8.19×10−6 cm2 s−1, respectively, calculating from the linear ip-v1/2 relationship (Figure S3B). The CV of Zn(OH)42−/Zn exhibits a different behavior on the electrode (Figures S3C and S3D), where both a surface-reaction-controlled process (linear dependence, ip∝v) and a diffusion-controlled process can be distinguished (Cheng et al., 2013). For better understanding the process of zinc deposition, a rotating disk electrode (RDE) technique was employed (Figure S4), demonstrating a diffusion-controlled process.

An alkaline zinc-iron flow battery is displayed in Figure S5. Solutions of 0.3 mol L−1 Zn(OH)42− and 0.6 mol L−1 K4Fe(CN)6 in 5 mol L−1 NaOH were used as negative and positive electrolytes, respectively. A self-made PBI membrane with an active area of 48 cm2 was employed in the structure. The battery was charged and discharged at a constant current density of 80 mA cm−2, affording a specific discharging capacity of 10.66 Ah L−1 with a starting discharging voltage of 1.82 V (Figure S6A). This high starting discharging voltage is unprecedented among all recently reported flow battery systems. The charge and discharge curves are given in Figure S6A. The battery exhibited an electrolyte utilization of 66% at a theoretical capacity of 16.07 Ah L−1, along with a coulombic efficiency (CE) of 99.97%, an EE of 88.07%, and a voltage efficiency (VE) of 88.11%. A permeability experiment of anionic active species through the PBI membrane was carried out to further verify the high selectivity of the membrane. As expected, the PBI membrane demonstrates an ultra-high selectivity for both positive and negative active species (Figure S7), which contributes to the high CE of the battery.

To confirm the reliability and practicality of the alkaline zinc-iron flow battery, the cycle performance was measured at a current density of 60 mA cm−2. An outstanding cycling stability with nearly 100% capacity retention can be obtained over more than 500 cycles with an over 99% CE and a nearly 90% EE (Figure S6B), indicating the high reversibility of Zn(OH)42− deposition/dissolution on the 3D porous carbon felt in the alkaline electrolyte.

Although excellent battery performance and cycling stability have been achieved, the concentration of active species in the supporting electrolyte is limited, leading to a low energy density. Thus to verify the practicability and afford a high-energy-density alkaline zinc-iron flow battery, further work was carried out to increase the active species concentration in both positive and negative electrolytes, which is highly desirable for practical application of a flow battery. Normally, K4Fe(CN)6 [or Na4Fe(CN)6] has a low solubility in KOH (or NaOH) solutions because of the “common ion effect.” Therefore to avoid the common ion effect, we opt to use sodium ferrocyanide as the positive active material, which affords a concentration of 1 mol L−1 in 3 mol L−1 potassium hydroxide supporting electrolyte (Xia et al., 2012), and 0.5 mol L−1 Zn(OH)42− dissolved in 4 mol L−1 sodium hydroxide as the negative electrolyte. The resulting flow battery with a self-made membrane demonstrated a well-defined OCV of 1.83 V at 50% states of charge (SOC). The OCV was monotonically increased from 1.74 to 1.96 V with SOC increasing from 5% to 90%, as evidenced in Figure S8. This OCV was higher than the estimated thermodynamic potential of 1.63 V, which was mainly attributed to the formation of cationic complexes with Fe(CN)63− and Fe(CN)64− ions (McBreen, 1984). The flow battery with a self-made membrane exhibited a discharge capacity and energy of ∼24 Ah L−1 and ∼40 Wh L−1, respectively, achieving an electrolyte utilization of ∼90% (theoretical capacity of 26.76 Ah L−1) at a current density of 60 mA cm−2 (Figure S9A). The CE, EE, and VE are 99.15%, 91.28%, and 92.06%, respectively, exhibiting a stable cycling performance for about 150 cycles (Figures S9B–S9D). By operating at a current density of 80 mA cm−2, a CE of 99.71%, an EE of 89.59%, and an electrolyte utilization of ∼80% can be achieved, as shown in Figure S10A, demonstrating a stable battery performance (Figures S10B and S10C) and a super-high-capacity retention (Figure S10D) for more than 200 cycles.

This excellent battery performance clearly validates the practicability of the alkaline zinc-iron flow battery. Nevertheless, the current density is relatively low, which may indeed affect the output power density of a battery. Therefore we further tested the alkaline zinc-iron flow battery by varying the current density from 60 to 160 mA cm−2. As expected, the battery still can deliver an outstanding battery performance, maintaining a CE of nearly 100% (Figures 2A and B). For instance, the battery exhibited no decrease in CE (∼100%) and EE (∼86%) after more than 200 cycles at a current density of 100 mA cm−2 (Figures 2A and 2C). It was also observed that this battery could be charged and discharged for about 40 min for each cycle while preserving a steady discharge capacity of above 21 Ah L−1 and discharge energy of above 34 Wh L−1 as the cycle proceeded (Figure 2C). Even at the high current density of 160 mA cm−2, the battery demonstrates an average VE of 83.20% and an average EE of 82.78% (Figure 2D), maintaining this performance over more than 150 cycles, together with a stable discharge capacity of 15.92 Ah L−1 and a discharge energy of 25.43 Wh L−1 (Figure 2D). Owing to the high CE of the battery, the discharge capacity of the battery remained nearly unchanged, yielding a capacity retention of over 99.999% per cycle. Such high performance at such high working current density is first reported among the newly developed flow battery systems. The battery shows comparable performance to a vanadium flow battery, although with a much lower cost and much higher power density owing to its higher OCV, showing promising prospect in large-scale energy storage.

Figure 2

Performance of the Alkaline Zinc-Iron Flow Battery Using a High-Concentration Electrolyte

(A) Cell voltage profiles of the alkaline zinc-iron flow battery with self-made PBI membrane at a current density of 100 mA cm−2.

(B) The performance of an alkaline zinc-iron flow battery with self-made membrane with current densities ranging from 60 mA cm−2 to 160 mA cm−2.

(C) The cycling performance of an alkaline zinc-iron flow battery at 100 mA cm−2. Insert, representative charge and discharge curves of the alkaline zinc-iron flow battery and the corresponding discharge capacity and discharge energy for each cycle.

(D) The cycling performance of an alkaline zinc-iron flow battery at 160 mA cm−2. Insert, representative charge and discharge profiles of the alkaline zinc-iron flow battery and the corresponding discharge capacity and discharge energy for each cycle. 60 mL 1.0 mol L−1 Na4Fe(CN)6 + 3 mol L−1 potassium hydroxide solution and 60 mL 0.5 mol L−1 Zn(OH)42− + 4 mol L−1 sodium hydroxide solution were used as the positive and negative electrolytes, respectively.

See also Figures S1, S5–S10, S13, S14, S17, and S21.

The Role of Membrane

To better understand the role of a PBI membrane, the performance of a battery assembled with a cation-conducting membrane (Nafion 115 membrane) was used for comparison. The surface and cross section of PBI (Figure S11) and Nafion 115 (Figure S12) membranes both demonstrated a smooth and dense structure. The distribution of negatively charged sulfonated groups of Nafion 115 can be further confirmed by transmission electron microscopy (Figure S12D). A battery assembled with a Nafion 115 membrane delivered a CE of 99.20%, an EE of 78.83%, and a VE of 79.46% at a current density of 80 mA cm−2 (Figure S13A), which is much lower than that delivered by a battery with a self-made PBI membrane. A battery with a Nafion 115 membrane shows a much higher initial charge-discharge voltage gap (145.1 mV) than a battery with a self-made membrane (49.78 mV) (Figure S13B), indicating a much higher ohmic polarization for the battery with a Nafion 115 membrane. The impedance Nyquist plot of a battery with a Nafion 115 membrane at 50% SOC demonstrates an ohmic resistance of 2.144 Ω cm2, which is nearly 2-fold higher than that of a battery with a self-made PBI membrane (1.152 Ω cm2) (Figure S14), further confirming the above-mentioned results. We propose that the poor performance of Nafion 115 membrane in this system originates from the repulsion effect between the negatively charged sulfonated groups in Nafion 115 (Figure S12D) and the negatively charged hydroxyl ions in the electrolyte, leading to poor ion conductivity. On the other hand, Nafion 115 has a more continuous phase-separated structure (Figures S15A and S15B) than a PBI membrane (Figures S15C and S15D). The hydrophilic phase from sulfonated acid groups of Nafion 115 results in a membrane with a swelling of 11.11%, which is much higher than that of a PBI membrane (4.348%). The continuous phase-separated structure together with high swelling of Nafion 115 may possibly result in zinc growth into the membrane easily during the charging process (Figure S16), further leading to high resistance of the membrane. By contrast, the PBI membrane is a totally aromatic rigid polymer and has very low swelling in aqueous solution, showing no obvious phase separation structure (Figures S15C and S15D); however, it can still easily transfer OH− by Grotthuss mechanism from imidazolyl groups. Taking the above-mentioned advantages into consideration, no obvious zinc dendrite and zinc accumulation, which are regarded as the two major issues hindering the commercial application of zinc-based batteries, could be observed on the carbon felt electrode by using a PBI membrane after charge and discharge cycling test, as shown in Figure S17. By contrast, the metallic zinc deposited on the electrode was found to be dense (Figures S18A–S18C) and obvious metallic zinc accumulation can be observed (Figures S18A′–S18C′) after cycling test by employing a Nafion 115 membrane. Taking the fast transportation of OH− through Grotthuss mechanism, no obvious phase separation structure, and the high mechanical stability of PBI together, a battery with a PBI membrane could retain a lower membrane resistance and higher VE than a Nafion 115 membrane, and the zinc dendrite and accumulation can be effectively prevented by artistically introjecting this membrane with our flow battery configuration.

The chemical stability of membrane materials during the charging-discharging process is one of the most critical concerns that affects the cycling stability of a flow battery system. Figure 3 shows the cycling stability of an alkaline zinc-iron flow battery using a self-made PBI membrane after treating it in 3 mol L−1 sodium hydroxide solution at 30°C for more than 1 month. The battery with the alkaline-treated membrane showed stable performance after continuously running for more than 500 cycles over a range of current densities (80–160 mA cm−2), suggesting that this kind of membrane can withstand the alkaline electrolyte solution in the long term (the battery runs 370 cycles at 80 mA cm−2 and 135 cycles at 160 mA cm−2. The duration of each cycle was 80 min for 80 mA cm−2 and 40 min for 160 mA cm−2. Thus the total time consumed for this test was 583 hr, which is long enough among the recently reported new flow battery systems). The surface and cross-sectional morphologies of the self-made PBI membrane after the flow battery cycling experiment were detected by field emission scanning electron microscope (FE-SEM) to further verify the membrane stability. As displayed in Figures S19A and S19B, similar to the pristine membrane (Figures S11A and S11B), a smooth and dense structure can be clearly observed, indicating that the membrane possesses excellent stability under alkaline medium. Besides, the membrane surface toward the positive half-cell remained flat and dense as well (Figure S19D), whereas the membrane surface facing the negative half-cell showed a relatively rough but still an integral morphology (Figure S19C), which was mainly attributed to the uneven plating/stripping of zinc during the charging and discharging process at high current density (160 mA cm−2). On the other hand, although [Fe(CN)6]3-/4- was reported to experience degradation in strong alkaline solution to some extent (Soloveichik, 2015, McBreen, 1984, Luo et al., 2017a), our long-term cycle experiment demonstrates that the degradation of [Fe(CN)6]3-/4- in our system is negligible and has little effect on the battery performance. Overall, the PBI membrane demonstrated an excellent stability in this alkaline zinc-iron flow battery medium.

Figure 3

Long-Term Stability of the Self-made Polybenzimidazole Membrane

180 mL 1.0 mol L-1 Na4Fe(CN)6 + 3 mol L−1 potassium hydroxide solution and 180 mL 0.5 mol L−1 Zn(OH)42− + 4 mol L−1 sodium hydroxide solution were used as the positive and negative electrolytes, respectively. The charging process was controlled by the charging time (40 min for 80 mA cm−2 and 20 min for 160 mA cm−2) to maintain the constant charge capacity, and the discharge process was ended with the cutoff voltage of 0.1 V. Thus the duration of each cycle was 80 min at 80 mA cm−2 and 40 min at 160 mA cm−2, respectively.

See also Figures S1, S7, S8, S14, and S19.

The Role of 3D Porous Carbon Felt Electrode

To investigate the role of a 3D porous carbon felt electrode, a battery with a PBI membrane using a zinc plate as the negative electrode (Figure S20A) was assembled. Shown in Figure S20 is the battery performance using a zinc plate electrode, affording a CE of 99.39% and an EE of 78.52% at a current density of 80 mA cm−2 (Figure S20B), which is much lower than that of a battery employing a 3D porous carbon felt as the electrode at the same condition. In addition, an uneven zinc plating (Figures S20C and S20E) and a distinctly severe zinc accumulation (Figures S20D and S20F) can be found for the battery using a zinc plate as the electrode at the end of charging and discharging. Even at a high CE of 99.39%, zinc accumulation still remained due to the poor connectivity between the deposited metallic zinc and the zinc plate electrode, further resulting in poor electronic conduction of the deposited metallic zinc. As a consequence, the Zn(OH)42− ions in the electrolyte were plated perpetually during charging, whereas metallic zinc on the zinc plate electrode was stripped during discharging, which eventually led to the zinc electrode being deformed and eroded. However, for a battery employing a 3D porous carbon felt with high specific surface area as the electrode, the zinc deposition (Figures S17A–S17C) and accumulation (Figures S17A′–S17C′) were significantly improved. Even with current densities increasing from 60 mA cm−2 to 160 mA cm−2, no significant difference in the morphologies of zinc deposited on the 3D porous carbon felt electrode could be observed, as shown by the results of scanning electron microscopy (Figure S21). This highly porous carbon felt considerably reduces the internal resistance of the interface between the electrode and deposited metallic zinc and prevents metallic zinc from falling (Figures S17A and S17B) before reaction because of the 3D porous structure. The high specific surface is propitious to further reduce polarization and improve the zinc deposition efficiency. During the charging process, the Zn(OH)42− was reduced and deposited on the porous carbon felt electrode, forming porous carbon felt/metallic zinc composite electrode (Figures S17B and S17C). Similar to previous reports (Luo et al., 2017b, Zhang et al., 2017), the pores of carbon felt could serve as a guidance on which the zinc stripping/plating processes can take place and effectively suppress the zinc dendrite/accumulation that occurs. During this process, OH− was released and transferred through the membrane to complete the internal circuit. The OH− in the negative electrolyte can go through the deposited metallic zinc and arrive at the positive half-cell in due time. Therefore the Zn(OH)42− in the bulk solution can be reduced and deposited on both the inner porous carbon felt and metallic zinc electrode. On the contrary, in the discharging process the metallic zinc was oxidized to Zn(OH)42− and OH− was consumed, which results in a concentration gradient between the positive and negative regions. Thus the fast transportation of OH− through the membrane can realize the fast oxidization of Zn on the porous carbon felt electrode, improving the accumulation of metallic zinc on the electrode. In addition, the deposition of metallic zinc on the porous carbon felt electrode was porous (Figure S14C). The porous metallic zinc can allow OH− to diffuse into the inner metallic zinc and electrode, which makes the metallic zinc deposited on the inner carbon felt discharged completely, as shown in Figure S17C’.

Practicability of the Alkaline Zinc-Iron Flow Battery System

Finally, to further confirm the practicability of the alkaline zinc-iron flow battery system, a kilowatt cell stack (Figure 4A) was assembled by using the self-made PBI membrane. Figures 4B–4D displayed the cell stack performance. The cell stack itself demonstrates an average CE of 98.84%, an average EE of 84.17%, and a VE of 85.16% at a current density of 80 mA cm−2, affording an average output power of 1.127 kW and an average discharge voltage of 16.10 V. Taking the advantages of the low-cost redox couples, self-made membrane, bipolar plate, and high voltage, a capital cost under $90 per kW h can be obtained (Tables S1–S6), which is much lower than the Department of Energy's cost target ($150 per kW h). To the best of our knowledge, this is the first time that a kW cell stack with a self-made, inexpensive membrane and such a low capital cost has been assembled and reported among the newly developed flow battery systems. Taken together, the excellent battery and cell stack performance (efficiencies and output power density) (Figures 5A and 5B), high energy density, and the super-low cost (Figure 5B) make the alkaline zinc-iron flow battery very promising for stationary energy storage.

Figure 4

Practical Realization of the Alkaline Zinc-Iron Flow Battery

(A) The kW alkaline zinc-iron flow battery cell stack prototype using self-made, low-cost non-fluorinated ion-exchange membrane.

(B) Cell stack voltage profile of the alkaline zinc-iron flow battery at a current density of 80 mA cm−2.

(C) Parts of charge and discharge curves of the cell stack.

(D) Cycle performance of the cell stack at a current density of 80 mA cm−2.

See also Figures S1 and S11.

Figure 5

Comparison of Both Recently Reported Flow Battery Systems and Traditional Flow Battery Systems with the Alkaline Zinc-Iron Flow Battery

(A) The performance of recently reported flow batteries. Most of these flow batteries use Nafion series membranes, which result in the high cost of the system.

(B) The cost and practical output power densities for different flow battery systems and the alkaline zinc-iron flow battery system. The practical output power densities were calculated from the charge-discharge curves of the single cells (the peak power density obtained from a polarization curve of a given flow battery was not considered here since at the point of the peak power density, the current density is normally much higher, e.g., higher than 600 mA cm−2; under this condition, the battery efficiency is too low, or the battery is difficult to work anymore). The error bars are defined as the cost fluctuations under different current density.

See also Figure S12 and Tables S1–S6.

Conclusion

In summary, we have demonstrated an ultra-high performance alkaline zinc-iron flow battery that can be operated at a wide range of current densities (60–160 mA cm−2). The battery exhibited very high power density, energy density, and efficiencies. Most importantly, by using the self-made, low-cost PBI membrane with ultra-high chemical stability, 3D porous carbon felt electrode, and inexpensive zinc and iron active materials, the cost of zinc/iron battery system is even lower than $90/kWh. The results indicated that the alkaline zinc-iron flow battery system is one of the most promising candidates for next-generation large-scale energy storage systems.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.