High Power Density Redox Flow Battery Cells. M. L. Perry, R. M. Darling, and R. Zaffou

Transcription

1 / ecst The Electrochemical Society High Power Density Redox Flow Battery Cells M. L. Perry, R. M. Darling, and R. Zaffou United Technologies Research Center, East Hartford, CT, 06108, USA Redox flow batteries possess several key advantages that make them well suited for grid-scale energy-storage applications. However, the capital cost of flow batteries has been a major barrier to commercialization of this technology. One attractive path to cost reduction is the development of flow-battery cells with substantially higher power densities than conventional flow-battery cells. The cost of the cells comprises a significant portion of the total flow-battery system cost, especially at low production volumes, since cell parts are custom-built components. UTRC has developed high power density vanadium-redox battery cells utilizing a material set similar to conventional cells. This advanced cell technology can theoretically be applied to other flow-battery chemistries as well. This breakthrough in cell performance should motivate additional development of flowbattery technologies, since a realistic path to substantial cost reduction, which can be realized even at low production volumes, has now been demonstrated. Introduction Major trends in both transportation and electric-grid technology are driving a growing demand for new Electrical Energy Storage (EES) solutions. In the transportation sector, a transition from mechanical drive to electric drive is ongoing with a series of vehicle technologies that rely increasingly on electric power for propulsion. Hybrid Electric Vehicles (HEV) being just the first step in this transformation to potentially the complete electrification of the automobile. On the electric grid, the intermittency of Renewable Energy Sources (RES), which have been becoming more prevalent, are often cited as the major motivation to incorporate substantially more EES. However, there are many other benefits to grid-scale EES, besides addressing the stochastic nature of RES, which is why some EES solutions are already being utilized to varying degrees in most electric grids (e.g., typically 2 to 15 % of electricity in developed countries is cycled through EES, with the vast majority being in the form of pumped hydroelectric power [1]). Essentially, since electricity demand is also highly cyclical, the grid can benefit from hybridization in a manner analogous to HEVs, even if all of the power sources on the grid are fully dispatchable and controllable. An electric grid with minimal EES has to be sized to accommodate the worst-case scenario, which results in poor utilization factors of both power generation and Transmission and Distribution (T&D) assets. Downloaded on to IP address. Redistribution 7 subject to ECS license or copyright; see ecsdl.org/site/terms_use

2 Historically, the relatively limited capabilities of conventional EES technologies have paced the implementation of energy storage in both transportation and grid applications. A major issue has been cost, both initial capital cost (i.e., $/kwh of rated energy capacity per discharge cycle) and lifecycle cost (e.g., the total Levelized Cost of Energy (LCOE) delivered over the lifetime of the EES system). Since these applications require at least several orders-of-magnitude larger energy-storage capacities than most portable applications (e.g., electronic devices such as cell phones or laptop computers), it is not surprising that the cost targets are much more aggressive. For example, most portable applications only require tens of watt-hours (Wh) of storage capacity or less, while transportation applications require kwh systems (e.g., 40 to 60 kwh for a AEV [1]) and many grid-scale applications require multi-mwh systems (e.g., on the order of 1 to 100 MWh for energy-time-shift applications [1]). Most conventional batteries cannot meet the aggressive cost targets cited for many grid-scale EES applications (e.g., < $500/kWh, installed, for even the most valuable of these applications [1]). And, even if cost targets can potentially be met with conventional batteries, the potential safety issues have considerably more serious implications with these large-scale EES systems. Additionally, cycle-life requirements for grid-scale EES are also typically substantially more aggressive than both portable and vehicle applications. Most grid applications require almost daily charge-discharge cycles over long lifetimes (e.g., typically 10 to 20 years, respectively), which means that thousands of potentially deep-discharge cycles is required. The strategy often employed to meet cycle-life requirements in other EES applications, which is to simply oversize the battery so that most cycles are relatively shallow (e.g., NiMH battery system in the Prius HEV), cannot be employed due to the aforementioned cost constraints. Redox Flow-Battery System Attributes A Redox Flow Battery (RFB) system is an Electrical Energy Storage (EES) approach that was originally conceived by NASA during the energy crises of the 1970s. Analogous to conventional secondary batteries, a RFB utilizes reversible electrochemical couples on two electrodes to store chemical energy. However, instead of storing the electrochemical reactants within the electrode, as in a conventional battery, the reactants are dissolved in electrolytic solutions and stored in tanks external to the reactor. The RFB reactor is a stack of cells; each cell contains sites where electrochemical chargetransfer reactions occur as the reactants flow through the cells. Therefore, the energy and power capacities of a RFB system are independent variables, unlike in conventional secondary batteries. A FBS can also undergo charge/discharge cycles without necessarily undergoing physical changes in the electrodes. When a conventional battery undergoes repeated charge/discharge cycles, the electrode materials typically expand and contract, which results in degradation over time and can limit the cyclic lifetimes of conventional batteries; this is especially true for batteries with relatively thick electrodes (i.e., high energy batteries) undergoing deep cycles. An excellent review article on RFB technology was recently published [2]. A FBS has inherent design flexibility that enables products with a wide range of powerto-energy ratios built upon common components or modules. (This is especially important for most stationary applications where each application tends to be unique and somewhat customized; therefore, the engineering time allotted per system is typically minimal relative to automotive applications, where the engineering costs can be Downloaded on to IP address. Redistribution 8 subject to ECS license or copyright; see ecsdl.org/site/terms_use

3 amortized over thousands, or even millions, of identical systems.) Additionally, flow batteries can have large energy capacities and high power capability in a single device. A FBS that offers both high energy and power capability can potentially be well suited for a wide range of EES applications. Although RFB systems are more complex than conventional batteries, this is less true for large EES applications, where a thermalmanagement system is often required with large battery packs and the total system complexity is therefore somewhat similar. In summary, among the myriad of battery technologies, RFB technologies possess several key advantages that make them particularly, if not uniquely, suited for Grid-Scale EES applications [3]. Despite all of these inherent advantages, RFB systems have not yet been successfully commercialized. The capital cost of a RFB system has been a major barrier to commercialization of this technology. Table I: A comparison between conventional and flow-cell batteries Technology Major attributes Major Issues Conventional rechargeable batteries (e.g., lead-acid, Li-ion) - High round-trip efficiencies - No precious-metal catalysts - High energy densities - Power & energy not independent - Limited life cycle - Continuous self-discharge Redox flow batteries (e.g., vanadium, polysulfide-bromide) - High round-trip efficiencies - No precious-metal catalysts - Energy and power are independent - Long life cycle - Low self-discharge rates - Complex (relative to conventional batteries) - Low energy densities (relative to conventional batteries) - Low power density cells Hybrid redox flow batteries (e.g., Zn-Br) - High round-trip efficiencies - Low self-discharge rates - Moderate energy densities - Complex (relative to conventional batteries) - Power & energy not independent - Limited life cycle (Zn electrode) - Precious-metal catalysts(br electrode) - Vapor pressure of Br 2 Path to Substantial Cost Reduction of Redox Flow-Battery Systems Cell stacks are a key module of any RFB system and also represents a substantial part of the total system cost. This is especially true at relatively low production volumes, since cell stacks are custom modules composed of custom-designed components (e.g., bipolar plates and electrode-membrane-seal assemblies), which are constructed from relatively costly materials (e.g., highly stable ion-exchange membranes and graphitic carbons). Therefore, substantial reductions in RFB cell-stack cost can result in significant RFB system cost reduction, which is required for successful commercialization of this technology. Downloaded on to IP address. Redistribution 9 subject to ECS license or copyright; see ecsdl.org/site/terms_use

4 Projecting costs of future systems is always an inherently challenging undertaking for multiple reasons, including: i) material costs are always subjected to dynamic market forces (e.g., supply-and-demand), ii) manufacturing costs per system are typically continuously decreasing (e.g., due to learning curves, which can be partially offset by increases in labor costs, which are also subject to market forces), and iii) there is always the possibility of discontinuities in product costs due to technology breakthroughs. In summary, cost metrics (e.g., $/kwh) are undoubtedly more nebulous than other Key Performance Indices (KPIs), which are primarily subject to only physical forces, not market forces (e.g., Wh/L or Wh/kg). Therefore, the approach taken here, with respect to cost analysis, is to simply show how one key physical KPI can impact system cost with a realistic, but broad, range of assumed component costs. The power density of a RFB cell is the KPI focused on here. This is defined as the power that can be stored and/or produced per a given active area of the cell (e.g., W/m2). Since the active area of each RFB cell typically consists of at least three major components: i) one separator (e.g., an ion-exchange membrane), ii) two electrodes (e.g., typically porous carbon media), and iii) one separator plate (e.g., typically a solid graphite plate), these are the key RFB-cell components that are considered to included what is henceforth referred to as cell repeat parts. The cost of these components, normalized for either the system s power capacity ($/kw) or energy capacity ($/kwh), is shown in Fig. 1. The assumed range of cell-repeat-part component costs chosen for Fig. 1 was $500 to $1,000/m 2, which is based on the fact that perfluorosulfonated (PFSA) membranes alone cost on the order of $500/m 2 when purchased in relatively low volumes. The carbon components are typically less costly than the membranes, but are certainly not insignificant. At low volumes, even molded-graphite-composite bipolar plates will typically cost in the range of $400/m 2, depending on both the size of the plates and the design features required. Fortunately, the electrode materials are the least expensive (e.g., on the order of < $100/m 2 at low volumes) since two electrodes per cell are needed. Some RFB systems may require or utilize cell repeat parts that are different than those listed above (e.g., may require metal catalysts in the electrode(s) or may utilize simple porous separators instead of ion-exchange membranes); however, even these variations in components may still fall in this range presented in Fig. 1 (i.e., but may be applicable at a different production volumes). Alternatively, Fig. 1 can readily be generated for any cost assumptions by the reader interested in doing so. It should be emphasized that the cell repeat parts considered here are not simply commodity materials that are purchased in bulk but, instead, are custom components that are therefore strong functions of production volume. Even if some of the raw materials are considered to be commodities because they are currently being produced for other applications (e.g., PFSA membranes are used in the chlor-alkali industry), the cost of these components includes fabricating parts that can actually be used in a RFB cell stack (whether this fabrication is done by the component suppliers and/or the RFB OEM). It should also be noted that other stack components are not considered here (e.g., manifolds, end plates, etc). Therefore, Fig. 1 does not account for the cost of a complete stack. The costs shown in Fig. 1 are for the cell repeat parts only, which do account for the majority of the stack components, but it also does not include the cost of assembling cell stacks. In summary, the range of costs considered here are based on realistic estimates for both the labor and materials required to make cell repeat parts that can be utilized directly in a Downloaded on to IP address. Redistribution 10 subject to ECS license or copyright; see ecsdl.org/site/terms_use

5 stack, with the only additional labor required being the assembly of the cell stack. This approach more accurately reflects the true cost of these components, without including complete stack costs, which would require too many assumptions about stack sizes and design features that would over complicate this simple analysis. Figure 1. Cost of RFB cell repeat part components as a function of the flow-cell power density. Left-hand axis is cost per power capability and right-hand axis is cost per energy capacity, assuming that RFB system capacity is 5-hr of rated power per discharge cycle. A typical rated-power operating point (100 mw/cm 2 of cell active area) for most RFB cells, to date, is also highlighted in the figure. The intent of Fig. 1 is to focus on a cost range that is realistic for low-to-moderate production volumes, i.e., on the order of 1,000 medium-size RFB systems produced per year (with medium-size systems defined as on the order of 1-MW with 4-6 MWh energy capacity per system). Since no RFB company is presently manufacturing 20 units/week of any size, these moderate production volumes cannot be immediately realized; therefore, the higher costs that result with even lower production volumes are of more immediate interest. However, arguably, the most important costs are moderateproduction costs, since market subsidies may exist that can support very low production volumes (e.g., prototype units and initial field demonstrations of different EES applications); however, such government-funded subsidies are usually not sufficient to support moderate production volumes (especially from multiple product developers). Cost projections at high production volumes (where the labor/material cost ratio is small and typically assumed to asymptote to a very small value at ever increasing volumes) are useful and necessary to understand what costs might ultimately be realized for a particular technology and, therefore, can help to identify what technologies may have Downloaded on to IP address. Redistribution 11 subject to ECS license or copyright; see ecsdl.org/site/terms_use

6 mass-market potential. Since material costs dominate such projections, the actual cell materials and chemistries being considered can be crucial. With lower production volumes, one can generically consider multiple types of RFB cells, which is one of the purposes of the simple approach presented here. Two key conclusions from Fig. 1 are i) increases in cell power density can have a significant impact on the cost of a commercially-viable RFB system and ii) the sensitivity to repeat-part component costs decrease substantially with higher cell power densities. The latter is especially important at low production volumes, since these costs tend to be a much stronger function of quantity of components ordered in this early production stage. The costs of the cell-repeat-part components may be prohibitive to enabling RFB systems that can meet commercial cost targets (e.g., installed system cost of ~ $500/kWh [1] or even much less 1 ), if cells operate at the power densities that have typically been reported to date (e.g., approximately 100 mw/cm 2, as shown below in Fig. 2). If one assumes that the energy capacity of the RFB system is 5-hrs at rated-power capacity 2, then a power density of 100 mw/cm 2 equates to $100 to 200/kWh for just the cell components, which represents a substantial portion of the total cost target for part of a single sub-system (i.e., just the cell components that are utilized in the stack). As noted above, RFB stack costs also must include the cost of assembly of the stack from these repeat-part components, plus the cost of the stack s non-repeat components (e.g., end plates, manifolds, etc.). Additionally, the RFB System costs will include several other major sub-systems, such as the reactants (i.e., positive- and negative-electrolyte solutions), the storage tanks, the power-electronics equipment, a control system, and the remaining balance-of-plant (e.g., pumps, valves, plumbing, etc.), along with the cost of assembling the system, as well as installation costs. High Power Density Redox Flow-Battery Cells United Technologies Research Center (UTRC) has developed RFB cells with substantially higher power densities than conventional RFB cells. An example is shown in Fig. 2, which compares one of UTRC s RFB cells with the performance of other Vanadium-Redox Battery (VRB) cells that had been reported in the literature prior to UTRC s recent work on RFB technology. The high power densities reported here can theoretically be obtained with any other flow-battery chemistries (i.e., no plating, or storage, of reactants in the electrodes), since one key factor to obtaining these high power densities is to design cells for maximum power, not to accommodate the storage of reactants (e.g., utilize relatively thin electrodes). UTRC s high power density cells utilize a set of cell materials that are similar to conventional RFB cells (i.e., carbon bipolar plates, porous-carbon electrodes, and ion-exchange membranes) and the results presented here were obtained utilizing commercially-available materials. The key to obtaining this substantially higher cell performance is improved cell designs; a complete description of which is beyond the scope of this Transactions paper. 1 The cost target for ARPA-E s GRIDS program is $100/kWh. 2 Some flow-battery developers prefer to quote costs for much longer discharge times (e.g., hrs) for obvious reasons; however, this discharge duration is not typically required and what should really be used for the purpose of estimating potential value propositions is the average-discharge time per cycle over the lifetime of the system, which is likely to be in the 4-to-6 hr range for most EES applications [4]. Downloaded on to IP address. Redistribution 12 subject to ECS license or copyright; see ecsdl.org/site/terms_use

7 Most VRB cells reported in the literature prior to 2009 were operated at relatively low current and power densities, i.e., approximately 100 ma/cm 2 and < 150 mw/cm 2, respectively. Two examples are presented in Fig. 2 for VRB cells, which span the performances reported over two decades for VRB cells. UTRC s cell technology enables substantially higher power densities, as shown in Fig. 2 where the peak power is > 1300 mw/cm 2 for UTRC s VRB cell compared to < 150 mw/cm 2 for previous cells. However, it is unlikely that UTRC s peak power density would actually be utilized in a practical RFB system due to the relatively low voltage efficiency associated with this result. Nevertheless, UTRC s cells generate relatively high power densities even at comparable voltage efficiencies. For example, at ~ 1.3 V UTRC s power density is > 500 mw/cm 2 (refer open arrow symbols in Fig. 2), which is considerably higher than conventional VRB cells. Note that the cost curves presented in Fig. 1 essentially asymptote at approximately 500 mw/cm 2, so this high cell power density being achieved at a useful voltage efficiency is definitely beneficial from a cell cost perspective. Figure 2. Discharge polarization curves (left-hand axis) and corresponding cell power densities (right-hand axis) of three different Vanadium-Redox Battery (VRB) cells. The first is an early VRB result [5], the second cell was the best VRB cell performance reported in the literature when UTRC began working on RFB technology [6], and the third is UTRC s cell performance. All were obtained at high states of charge (> 80%). Even though the peak power density is obtained at relatively low voltage efficiency, this high current-density performance is still beneficial, since it results in cell performance that does not vary much with state-of-charge at higher cell voltages. In other words, UTRC s RFB cell design enables high limiting currents since the masstransport losses have been minimized, which means that as the concentration of the active species decreases with state-of-charge the corresponding impact on the cell performance Downloaded on to IP address. Redistribution 13 subject to ECS license or copyright; see ecsdl.org/site/terms_use

8 is also minimized. Additionally, any increases in the mass-transport losses (e.g., due to cell degradation with time) should have less of an impact on the performance of the cells, especially in the operating range of primary interest. Furthermore, UTRC s cell performance potentially enables RFB systems with broad range of power outputs, albeit at varying energy efficiencies, which may be useful in some EES applications (e.g., capital deferral EES applications [4]). High power density cells are a previously unexploited advantage of RFBs; namely, RFB cells can be designed to operate at much higher current and power densities than conventional secondary batteries, which is another benefit enabled by the inherent power and energy independence of the RFB architecture. Therefore, in addition to being capable of storing a large amount of active material relative to inactive materials (i.e., high energy-to-power ratios), the RFB architecture also enables a further reduction in the amount of inactive materials required by utilizing high power density cells. Maturation of High Power Density Redox Flow-Battery Cells As part of UTRC s GRIDS project supported in-part by DOE s ARPA-E Office, the high power density cell technology demonstrated above has been scaled up to large cells (> 800-cm 2 of active area per cell). Large cell stacks with these full-size cells have then been assembled and tested in a complete RFB system at UTRC. UTRC s major deliverable for the GRIDS project was an Advanced Prototype System (APS), which has a single stack that can generate or store 20-kW of power. The tanks are large enough to enable this system to run for more than 1-hr at rated power per charge or discharge cycle and for longer at lower power conditions where the system efficiency is higher. Like any true RFB system, the energy capacity can be increased by simply adding more tank capacity, without impacting the rest of the system. In fact, UTRC is presently building a Field Demonstration System (FDS) with approximately 100-kWh energy capacity and similar range of power capacities as the APS. This FDS will be utilized to further mature this promising technology by operating this system in real-world operating conditions with realistic load profiles, to the greatest extent possible with this size of system. The FDS is packaged in a standard 20-foot shipping container to enable simplified deployment and installation of this demonstration system at locations other than UTRC. Path Forward Other groups have also begun to report on the development of high power density RFB cells, such as Lawrence Berkeley National Lab (LBNL) and the University of Kansas with Hydrogen-Bromine cells [7, 8], and the University of Tennessee, Knoxville with VRB cells [9]. Ideally, this trend should result in cell-component suppliers optimizing RFB cell components for high power density operations. To date, these higher cell power densities have been obtained using existing commercial components, and even higher performances can be anticipated with advanced materials. In fact, there has been very limited RFB technology development over the past three decades, presumably in part because there was not a realistic path for this technology to achieve commercially-viable capital costs. There are many potential opportunities to further improve these technologies, beyond advanced cell components (e.g., advanced chemistries, system design and control strategies, etc.). Ideally, an entire technical community working on Downloaded on to IP address. Redistribution 14 subject to ECS license or copyright; see ecsdl.org/site/terms_use

9 these technologies with a development path that is firmly grounded in scientific advances will result in future technical breakthroughs that can make these technologies increasingly attractive. A growing program in applied research, development, and demonstration to examine many of the common and synergistic RFB components and phenomena is still required, many of these promising research opportunities have been summarized elsewhere [10]. Conclusions RFB cells with substantially higher power densities than conventional flow-battery cells have been developed. This breakthrough in cell performance has also been successfully scaled-up to large-format cells, which have been built into cell stacks and tested in a complete Advanced Prototype System at UTRC. High power density flow-battery cells take advantage of the inherent power and energy independence of the RFB architecture in a manner that had not been previously exploited. Namely, flow-battery cells can be designed for maximum power, independent of the quantity of reactants stored in the tanks of a RFB system. As shown here, RFB cells can be designed to operate at much higher current and power densities than conventional secondary batteries, which will further add to the potential attractiveness of RFB technologies for many grid-scale EES applications. In particular, high power density RFB cells enable a realistic path to substantial cost reduction of the complete system, which can be realized even at low production volumes where achieving cost targets is particularly challenging. Additional future improvements in RFB technology should be expected, especially if more researchers with fuel-cell expertise work in this area, since RFB technology has not received a lot of R&D attention to date (i.e., compared to fuel cells or conventional batteries) and much of the know-how required to developed improved fuel-cell systems can potentially be applied here as well. Acknowledgments The authors would like to thank their flow-battery project colleagues at UTRC. The work presented herein was funded, in part, by the Advanced Research Projects Agency - Energy (ARPA-E), U.S. Department of Energy (DOE) under Award Number DE- AR The information, data, or work presented herein was funded in part by an agency of the United States Government. Neither the United States Government nor any agency thereof, nor any of their employees, makes any warranty, express or implied, or assumes any legal liability or responsibility for the accuracy, completeness, or usefulness of any information, apparatus, product, or process disclosed, or represents that its use would not infringe privately owned rights. Reference herein to any specific commercial product, process, or service by trade name, trademark, manufacturer, or otherwise does not necessarily constitute or imply its endorsement, recommendation, or favoring by the United States Government or any agency thereof. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof. Downloaded on to IP address. Redistribution 15 subject to ECS license or copyright; see ecsdl.org/site/terms_use

10 References 1. Report on the First Quadrennial Technology Review: Technology Assessments. DOE/S- 0002, U.S. Department of Energy, Washington, DC. (2012). 2. A. Z. Weber, Ma. M. Mench, J. P. Meyers, P. N. Ross, J. T. Gostick, and Q. Liu, Redox flow batteries: a review Journal of Applied Electrochemistry, 41, pp (2011). 3. Z. Yang, J. Zhang, M.C.W. Kintner-Meyer, X. Lu, D. Choi, J.P. Lemmon, and J. Liu, Electrochemical Energy Storage for Green Grid, Chem. Reviews 111, pp (2011). 4. J. Eyer and G. Corey, Energy Storage for the Electricity Grid: Benefits and Market Potential Assessment Guide, SANDIA Report (2010). 5. M. Kazacos and M. Skyllas-Kazacos, Performance characteristics of carbon plastic electrodes in the all-vanadium redox cell, Journal of the Electrochemical Society, 136 pp (1989) 6. P. Qian, H. Zhang, J. Chen, Y. Wen, Q. Luo, Z. Liu, D. You, and B. Yi, A novel electrode-bipolar plate assembly for vanadium redox flow battery applications, Journal of Power Sources, 175, pp (2008). 7. K. T. Cho, P. Ridgway, A. Z. Weber, S. Haussener, V. Battaglia, and V. Srinivasan, High performance hydrogen/bromine redox flow battery for grid-scale energy storage, Journal of the Electrochemical Society, 159, pp. A1806-A1815 (2012). 8. T. V. Nguyen and H. Kreutzer, Effect of Transport on the Performance of a Hydrogen Bromine Flow Battery, ECS Transactions, 41, No. 11, pp. 3-9 (2011) 9. D. S. Aaron, Q. Liu, Z. Tang, G. M. Grim, A. B. Papandrew, A. Turhan, T. Zawodzinski, and M. Mench, Dramatic performance gains in vanadium redox flow batteries through modified cell architecture, Journal of Power Sources, 206 pp (2012). 10. Flow cells for Energy Storage, Workshop Summary Report (2012); online at: Downloaded on to IP address. Redistribution 16 subject to ECS license or copyright; see ecsdl.org/site/terms_use