Stationary electric energy storage is anticipated to play an increasingly important role in the efficient, reliable, and sustainable delivery of electricity, especially with increasing deployment of low cost intermittent renewable energy sources. Redox flow batteries (RFBs) are a promising electrochemical technology whose decoupling of power and energy scaling, long operational lifetimes, and safety are particularly appealing for energy storage with long discharge duration. The recent emergence of aqueous organic redox flow batteries (AORFBs) offers intriguing new pathways to inexpensive energy storage through the use of charge storage materials that are composed of earth-abundant elements and have the potential for cost-effective mass production, and whose electrochemical and physicochemical properties can be tuned through molecular functionalization.(1−3) Further, redox-active organics possess several secondary benefits, including compatibility across a wide pH range, rapid redox kinetics with multielectron transfer, and low permeability through polymeric membranes. Over the past decade, many organic and organometallic molecular families have been proposed and pursued, including but not limited to quinones,(1,4−10) viologens,(11−13) nitroxide radicals,(11,12,14−18) aza-aromatics,(19−24) and iron coordination complexes.(13,25) In addition, several start-up companies have emerged seeking to translate these scientific advances into cost-competitive energy solutions, such as Kemiwatt (France), Green Energy Storage (Italy), XL Batteries (United States), and JenaBatteries and CMBlu (Germany). Despite this progress, molecular lifetime remains a significant challenge for practical deployment for AORFBs.(13,26) Initial durability experiments tend to involve repetitive galvanostatic charge–discharge cycles, typically in a bulk electrolysis cell, at low concentrations and/or small electrolyte volumes to reduce the materials- and time-intensity of experiments. Thus, early claims of long cycle life are often based on a large number of cycles without significant decay performed over a short period of time. With recent advances in the field, it has become evident that all reasonably long-lived molecular redox couples (here, defined as having a half-life greater than a day) of which we are aware, when sufficiently thorough experiments to distinguish between cycle- and calendar-denominated fade have been performed, have shown decay to be time-dependent and a function of temperature, state of charge (SOC), molecular concentration, pH, and electrostatic potential.(7,8,13,26) Generally, albeit with some exceptions,(7) the molecule is more suspectible to decay in its “energized” state (i.e., the oxidized state for the posolyte or the reduced state for the negolyte). The continued development of AORFBs requires the adoption of more rigorous testing protocols to accurately assess molecular lifetimes, to elucidate the subtle failure modes of increasingly robust molecules, and to maintain the iterative design–test–improve cycle. Moreover, as the cost-effectiveness of AORFBs is dependent on the chemical costs and fade rates of finite-lifetime organics, as opposed to the current state-of-the-art vanadium, we aim to help guide and accelerate research efforts by articulating a feasible design space for the community to target. In this Viewpoint, we first outline experimental protocols that we urge those active in the field to adopt to better quantify capacity fade. Subsequently, we extend an existing bottom-up capital cost model to estimate AORFB system cost as a function of molecular lifetime, chemical costs, and the interest rate for discounting. While we limit our focus to aqueous-soluble organics, we anticipate the following approaches are applicable to all molecular redox-active materials for which lifetime is uncertain. Capacity Fade Rate Protocol. The procedures summarized below are recommended because they improve the accuracy and simplify the interpretation of capacity fade rates due to molecular decomposition. It is encumbent upon the researcher to determine how closely to follow these procedures, based on prior evidence, the value of time in a study, or whether certain analyses should be expanded upon to address chemistry-dependent phenomena. We note that these procedures are representative and intended as guidance to the greater community. There is room and need for continued innovation and advancement in this discipline. Cell Configuration. A volumetrically unbalanced, compositionally symmetric flow cell configuration,(26,27) in which the same compound fills both reservoirs at an intermediate state of charge (SOC), is recommended for characterizing capacity fade rates (Figure 1a). The symmetric flow cell technique offers a controlled electrolyte environment by removing the need for a counter electrode of dissimilar material, suppressing the loss of capacity through active species crossover. Note, however, that if the oxidized and reduced forms of a species have different membrane permeabilities and the SOC history of the capacity-limiting side is not managed well, then the possibility of net crossover exists even in a symmetric cell. Symmetric compositions also effectively eliminate the possibility of side-product species crossing over from the counter electrode chamber and contaminating the working electrode. The flowing electrolyte improves mass transfer, enabling cycling of concentrated solutions and investigatation of active material stability on high surface area porous electrodes of relevance to practical flow battery applications. Care should be taken when selecting wetted components, and ex situ compatibility studies are recommended prior to experimentation. The charge capacity of the electrolyte of interest should first be evaluated with in situ or ex situ bulk electrolysis, with discussion given to any significant discrepancy from theoretical capacity. The volumes filling the reservoirs should differ so that one side is the capacity-limiting side (CLS) during both oxidation and reduction. The noncapacity-limiting side (NCLS) should have active species in significant excess so that it remains noncapacity-limiting in both oxidation and reduction, even if undesirable side reactions act to drive the cell away from perfect electrolyte balance. If the active species of interest is confined to the CLS and the crossover rates of the active and contaminant species through the membrane are sufficiently low, this methodology can be applied to full cells as well. As a working hypothesis, it is reasonable to assume that the decomposition rates of the individual active species will not change and that the major effect of switching to full-cell configurations would be the additive effect of crossover. However, this hypothesis may be challenged by various effects including differences in the potential gradients within the cell and chemistry-specific interactions between constituent components. Figure 1. (a) Schematic of volumetrically unbalanced compositionally symmetric cell, with identical electrolytes at identical concentrations at 50% SOC in reservoirs of different volumes. Active species crossover is suppressed, and measured capacity fade may be attributed to the state of the capacity limiting side (CLS). (b) Potentiostatic cycling with polarity switching occurring when the current density drops below 1 mA/cm². From ref (26). Copyright 2018. Cycling Methodology. A potentiostatic hold at the end of a charging step and of a subsequent discharging step is essential for accurate capacity measurements. This can be implemented through purely potentiostatic cycling, as in Figure 1b, with the voltage held until the current density drops to very near an empirically determined background value, typically of order 1 mA/cm². Galvanostatic cycling, the more common approach to charge–discharge cycling, is vulnerable to artifacts that evolve over time during cell operation and can obscure low capacity fade rates, as shown in Figure 2. While both techniques require judicious selection of voltage limits, often established through a combination of voltammetric and mass transport analysis, with constant current operation drifts in the internal resistance of the cell, e.g. from diurnal temperature swings, from aging of the wetted components, and even from pulsatile flow at low SOC, can appear as apparent capacity fade. Alternatively, galvanostatic cycling with a potential hold at the end of each half-cycle is an effective means of eliminating these artifacts and isolating the true capacity fade rate (Figure 2). If, for whatever reason, it is necessary to evaluate the capacity fade rate during purely galvanostatic cycling, it is recommend that one in every nth cycle is followed by a potential hold in the charging and discharging half-cycles and that the discharge capacities from only these cycles be used to assess capacity fade rate. The inclusion of pauses at various SOCs can enable assessment of SOC-dependent temporal fade rates upon the resumption of cycling, as illustrated in Figure 3. Systematic variations in cycle period can, in principle, uniquely disaggregate cycle-dominated fade rates from time-dominated fade rates; however, as Figure 3 shows, the instantaneous fade rate may also depend on the remaining capacity. Figure 2. Semilog plot of volumetrically unbalanced compositionally symmetric cell cycling of 2,6-dihydroxyanthraquinone (DHAQ), showing changes in apparent capacity with adjustments to overall cell resistance and current density for purely galvanostatic cycling and the absence of such artifacts under potentiostatic cycling and in galvanostatic cycling with potential holds at the end of each half-cycle. In the latter two cases, cycling was performed by imposing voltage holds at ±200 mV and switched when current density dropped to 1 mA/cm². Galvanostatic conditions were started at 10 mA/cm² and increased to 20 mA/cm² after 2.13 days. Vertical arrows indicate times at which a resistor (130 mΩ, equivalent to a 25% increase in ohmic resistance) was added and removed in series with the cell. The jumps in apparent capacity demonstrate the dependence of apparent cycling capacity on cell resistance when strictly galvanostatic conditions are used. From ref (26). Copyright 2018. Figure 3. Semilog plot of volumetrically unbalanced compositionally symmetric cell cycling of 50% SOC 0.1 M DHAQ in 1 M KOH with cycling pauses in different states of charge. Instantaneous decay rates derived from slopes of each cycling segment are denoted in black text. Decay rates during holding periods at various SOCs are denoted in blue text. Every 20th cycle is denoted by a tick mark along the top axis. Adapted from ref (26). Copyright 2018. Data Interpretation. Only after the capacity has faded to a value measurably below the capacity measured originally in a bulk electrolysis experiment can one begin to interpret capacity fade rate measurements quantitatively. Before that, it is unclear whether the apparent capacity is being maintained by recruiting species that, for whatever reason, were not accessed during the first few cycles. For example, the presence of residual dissolved O₂ in the electrolyte at the start of cycling experiments may disturb initial cycles by chemically oxidizing a reduced negative electrolyte. If it is suspected that the NCLS may have become capacity-limiting during the cycling experiment, at the end of cycling the NCLS should be replaced with fresh electrolyte and a few additional full charge–discharge cycles imposed to confirm the final capacity of the CLS. If a cycling regimen involves a restricted SOC swing, its capacity fade rate should be evaluated by comparing full charge–discharge cycles immediately before and after the restricted SOC regimen: otherwise, it is unclear whether the apparent capacity is being maintained by recruiting species that were not accessed during the restricted SOC swing regimen. Should capacity fade rates be too low to measure with confidence, it may be helpful to accelerate reactions by increasing temperature, either of electrolytes in isolation or within an operating cell. For example, it is straightforward, though tedious, to hold both the oxidized and reduced forms (the latter may need to be very well protected from air, as described in the Supporting Information in ref (8)) at elevated temperature for multiple days and perform extensive chemical analyses (e.g., mass spectrometry, NMR, and cyclic voltammetry). However, caution should be taken as the dominant mechanism may be temperature-dependent; additionally, redox-active and inactive decomposition products must be distinguished. Furthermore, active species lifetimes may be different in isolation than in contact with wetted materials in an operating cell. Data Reporting. It is recommended that both capacity fade rate per day and per cycle be reported. If, as is typical, a certain fractional capacity loss is observed but not uniquely decomposed into cycle- and time-denominated rates, then the capacity fade rate can be reported as “X% per day or Y% per cycle”. Capacity fade rate rather than capacity retention rate should be reported because the arithmetic for converting between hourly, daily, and annual fade rates is simpler than that for converting between the corresponding capacity retention rates. It is important to report sufficient information regarding the experimental methods, including electrolyte volumes, to enable a knowledgeable colleague to reproduce the experimental results. Cost–Lifetime Trade-off. Under certain scenarios, molecular lifetime does not have to be infinite for an organic chemistry to be potentially lower in cost than the incumbent vanadium chemistry. Indeed, the up-front capital cost savings of using an inexpensive organic instead of vanadium can be compared with the present value of a series of future replacement costs, recognizing the time-value of money. The trade-off is quantified by a replacement cost ratio, defined as the annual replacement cost divided by the up-front capital cost savings.(2) The break-even value of the replacement cost ratio depends on the interest rate for discounting and the project lifetime, as shown in Figure 4. When the actual value of the replacement cost ratio is less than this break-even value, organics are potentially lower-cost than vanadium. For example, if the organic is 40% of the cost of vanadium per kilowatt-hour but loses 9% per year, then the replacement cost ratio is (0.09)(0.4)/(1.0 – 0.4) = 0.06 and a 20-year project breaks even at an interest rate of 1.8%, with higher interest rates favoring the organic. If one anticipates an infinite-year project for vanadium, enabled by periodic electrolyte maintenance, then under the same assumptions as before, break even would occur at an interest rate of 6.0%. This simple comparison assumes zero reconditioning cost when organic or vanadium replacements occur, but the model can be generalized by adding reconditioning cost. Figure 4. Break-even value of replacement cost ratio vs interest rate for discounting, assuming project lifetimes as indicated. Adapted from ref (2). Copyright 2018. Design Space for Electrolytes of Finite Lifetime. The effects of finite electrolyte lifetime and necessary replacements can be integrated into existing capital cost models to account for the present value of a series of future chemical replacement costs and to outline the allowable ranges of chemical costs, fade rates, and discounting rates and their dependence on other relevant cost-constraining parameters.(28,29) For brevity, the mathematical manipulations, simplifying assumptions, and baseline model inputs are detailed in the Supporting Information. Here, we illustrate the impact of periodic electrolyte replacement on the economic viability of AORFBs as a function of capacity fade rate, discount rate, and discharge duration and properties of the redox chemistry (e.g., cell potential, chemical costs, and cell resistance) (Figure 5). As shown in Figure 5a, AORFBs with high cell voltages and low total active costs have reduced up-front capital costs, especially at longer discharge durations (10–100 h) where the system cost is dominated by the electrolyte materials.(30) For comparison, the vanadium RFB capital costs, assuming no electrolyte maintenance or replacement over the project lifetime, are shown as a function of vanadium pentoxide precursor prices (average ± standard deviation) based on publicly available data from the past three years.(31) For reference, capital costs of vanadium RFBs are frequently cited between 300 and 500 $/kWh, and our calculated values fall within this range at comparable storage durations (1–10 h).(32−34) The variability in vanadium price, dictated by other industries (e.g., steel manufacturing), offers a window for finite-lifetime organics provided the redox chemistries have suitable combinations of high cell voltages and low electrolyte costs. The latter may be realized through a balance of inexpensive chemicals, low fade rates, and favorable interest rates for discounting (Figure 5b). Visualizing the dependence and sensitivity of capital cost to total active cost as a function of cell potential, cell resistance, and duration also provides insight into the relative importance of reactor performance and cost in defining cost-effective applications spaces (Figure 5c,d). Figure 5. Capital cost modeling of generic RFBs incorporating the net present value of the finite-life electrolytes. (a) Capital cost decreases with longer discharge duration, higher power reactors, and inexpensive electrolytes. At long durations (10–100 h) the system costs begin to asymptotically approach that cost of the electrolyte in the reservoirs. The cost of vanadium RFBs for a range of recent historical vanadium prices is provided as a comparison. (b) The total active costs, which is directly proportional to the electrolyte cost, is a function of the chemical cost, the fade rate, and the discount rate. Capital cost vs total active cost for various discharge durations and (c) cell potentials; (d) area-specific resistances (ASRs). At low total active costs, the capital cost of the cell shows greater sensivity to cell potential, discharge duration, and cell resistance. At high total active costs, the electrolyte dominates the capital cost and, with an exception of cell voltage which impacts both power and energy costs, there is little differentiation between the other parameters. For the data shown in panels a–c, a cell ASR of 1.0 Ω·cm² is assumed, and for panel d, a cell voltage of 1.0 V is assumed. Further details on model inputs, formulation, and data visualization can be found in the Supporting Information. Summary. The cost of electrolyte materials for a RFB is a dominant factor for systems with high E/P ratios. For aqueous systems, in which the solvent and salt are generally inexpensive, the cost of the active materials is of greatest concern. Organic active materials, made of earth-abundant elements, can be, but are not necessarily, low-cost after synthesis. However, their finite lifetimes raise their costs by adding a periodic replacement cost that complicates techno-economic assessments. We have described the best methods available to us at this time on how to assess the lifetime of organic active materials. We have shown how to factor the lifetime into an estimate of the system cost per kilowatt-hour of energy storage capacity. The total active cost is directly proportional to the capital cost per kilowatt-hour of the active materials and is a function of the annual replacement ratio and the interest rate for discounting future expenses to the present. The cost of incumbent vanadium RFBs serves as a convenient foil for comparison. These results begin to define a design space for cost-competitive AORFBs that necessitates materials of sufficiently disparate redox potentials that are stable and economical. Of the numerous and expanding variety of organic redox couples reported in the literature,(35) a few redox chemistries with high cell potential (≥1 V) and stability (ca. 5% per annum or less)(8) can be identified, although, to the authors’ knowledge, none has been tested for durations approaching a year. In parallel, recent initial process economic assessments by Yang et al. and Dieterich et al. have shown potential pathways to low-cost materials (ca. $35/kAh per side, or less),(2,3) but those materials are often associated with higher fade rates. As the field progresses, further work is needed to articulate the scientific and economic trade-offs between materials cost, stability, and ease of recovery. To this last point, recent work has demonstrated the potential for in situ regeneration,(36) which may eventually serve as a capacity recovery mechanism similar to electrolyte rebalancing in vanadium RFBs, further improving the prospects of potentially inexpensive materials with moderate fade rates. Ultimately, the authors hope that this Viewpoint provides guidance to the research community and inspires new research directions that further advance the science and engineering of AORFBs. The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsenergylett.0c00140.

• Mathematical formulations, model inputs, and simplifying assumptions regarding the techno-economic analyses performed in this work (PDF)

Mathematical formulations, model inputs, and simplifying assumptions regarding the techno-economic analyses performed in this work (PDF) M.J.A.: conceptualization, methodology, writing–original draft, supervision, funding acquisition; FRB: methodology, writing–original draft, resources, supervision, funding acquisition; KER: methodology, software, investigation, visualization. Views expressed in this Viewpoint are those of the authors and not necessarily the views of the ACS. The authors declare no competing financial interest. Electronic Supporting Information files are available without a subscription to ACS Web Editions. The American Chemical Society holds a copyright ownership interest in any copyrightable Supporting Information. Files available from the ACS website may be downloaded for personal use only. Users are not otherwise permitted to reproduce, republish, redistribute, or sell any Supporting Information from the ACS website, either in whole or in part, in either machine-readable form or any other form without permission from the American Chemical Society. For permission to reproduce, republish and redistribute this material, requesters must process their own requests via the RightsLink permission system. Information about how to use the RightsLink permission system can be found at http://pubs.acs.org/page/copyright/permissions.html. Work at Harvard was supported by U.S. DOE award DE-AC05-76RL01830 through PNNL subcontract 428977, by ARPA-E award DE-AR0000767, by Innovation Fund Denmark via the Grand Solutions project “ORBATS” file nr. 7046-00018B, and by the Massachusetts Clean Energy Technology Center. Work at MIT was supported by the Joint Center for Energy Storage Research (JCESR), an Energy Innovation Hub funded by the United States Department of Energy, Office of Science, Basic Energy Sciences (DE-AC02-06CH11357). The authors thank Eric Fell, Martin Jin, Yan Jing, Emily Kerr, Daniel Pollack, Zhijiang Tang, Andrew Wong, and Min Wu (Harvard) as well as Michael Orella (MIT) for helpful comments. This article references 36 other publications.

中文翻译：

---

水性液流用氧化还原活性有机物的寿命和成本

预计固定式电能存储在高效，可靠和可持续的电力输送中将发挥越来越重要的作用，尤其是随着低成本间歇性可再生能源的部署不断增加。氧化还原液流电池（RFB）是一种很有前途的电化学技术，其功率和能量缩放比例的解耦，较长的使用寿命和安全性特别适合具有较长放电时间的储能。水性有机氧化还原液流电池（AORFB）的近来出现，通过使用由富含地球元素的电荷存储材料提供了引人入胜的廉价储能新途径，这些材料可能具有成本效益的大规模生产，（1-3）此外，氧化还原活性有机物还具有一些次要优点，包括在宽pH范围内具有相容性，具有多电子转移的快速氧化还原动力学以及通过聚合物膜的低渗透性。在过去的十年中，已经提出并追求了许多有机和有机金属分子家族，包括但不限于醌，（1,4-10）紫精，（11-13）氮氧自由基，（11,12,14-18）氮杂芳烃，（19-24）和铁配合物。（13，25）。此外，一些新兴公司正在寻求将这些科学进展转化为具有成本竞争力的能源解决方案，例如Kemiwatt（法国），Green储能（意大利），XL电池（美国）以及JenaBatteries和CMBlu（德国）。尽管取得了这一进展，但分子寿命仍然是AORFB实际应用的重大挑战。（13,26）最初的耐久性实验往往涉及重复的恒电流充放电循环，通常是在大容量电解槽中，低浓度和/或小电解液体积下进行的。减少实验的材料和时间强度。因此，长寿命的早期主张通常是基于大量的周期，而在短时间内没有明显的衰减。随着该领域的最新进展，很明显，当进行充分彻底的实验以区分循环-时，我们知道所有合理的长寿命分子氧化还原对（此处定义为半衰期大于一天）。并已执行以日历命名的淡入淡出，已经表明衰变是随时间变化的，并且是温度，电荷状态（SOC），分子浓度，pH和静电势的函数。（7,8,13,26）通常，尽管有一些例外，（7）分子更可能在其“激发”状态（即，对正电解质的氧化态或对正电解质的还原态）衰变。AORFB的不断发展要求采用更严格的测试方案来准确评估分子寿命，阐明日益坚固的分子的细微故障模式，并保持迭代的设计-测试-改进周期。此外，由于AORFB的成本效益取决于有限寿命有机物的化学成本和褪色速率，而与目前最先进的钒相反，我们的目标是为社区提供明确的可行设计空间，以帮助指导和加速研究工作。在此观点中，我们首先概述了实验方案，我们敦促那些活跃于现场的人员采用以更好地量化容量衰减。随后，我们扩展了现有的自下而上的资本成本模型，以估计AORFB系统成本为分子寿命，化学成本和折现利率的函数。尽管我们将注意力集中在水溶性有机物上，但我们预计以下方法适用于寿命不确定的所有分子氧化还原活性材料。我们扩展了现有的自下而上的资本成本模型，以估计AORFB系统成本作为分子寿命，化学成本和折现利率的函数。尽管我们将注意力集中在水溶性有机物上，但我们预计以下方法适用于寿命不确定的所有分子氧化还原活性材料。我们扩展了现有的自下而上的资本成本模型，以估计AORFB系统成本为分子寿命，化学成本和折现利率的函数。尽管我们将注意力集中在水溶性有机物上，但我们预计以下方法适用于寿命不确定的所有分子氧化还原活性材料。容量衰减率协议。建议使用以下总结的程序，因为它们可以提高精度并简化由于分子分解而引起的容量衰减率的解释。研究人员有责任根据先前的证据，研究的时间价值或是否应扩大某些分析范围以解决化学依赖性现象，来确定遵循这些程序的紧密程度。我们注意到，这些程序具有代表性，旨在为更大的社区提供指导。在这个学科上有继续创新和进步的空间和需求。单元配置。建议使用体积不平衡，组成对称的流通池配置（26,27），其中相同的化合物以中间电荷状态（SOC）填充两个储层，以表征容量衰减率（图1a）。对称流通池技术通过消除对异种材料的反电极的需求，从而抑制了由于活性物质交叉而导致的容量损失，从而提供了受控的电解质环境。但是请注意，如果某个物种的氧化形式和还原形式具有不同的膜渗透性，并且不能很好地控制容量限制侧的SOC历史，那么即使在对称电池中也存在净交叉的可能性。对称的成分还可以有效消除副产物从对电极室越过并污染工作电极的可能性。流动的电解液改善了质量传递，使浓溶液循环，并研究了与实际液流电池应用相关的高表面积多孔电极上活性物质的稳定性。选择浸湿的组件时应小心，并且建议在实验前进行异位相容性研究。首先应通过原位或异位评估感兴趣的电解质的电荷容量本体电解，讨论与理论容量之间的任何重大差异。填充容器的体积应有所不同，以使一侧在氧化和还原过程中均是容量限制侧（CLS）。非容量限制的一面（NCLS）应具有显着过量的活性物质，以使它在氧化和还原方面仍保持非容量的限制，即使不希望的副反应使电池脱离了完美的电解质平衡。如果感兴趣的活性物质仅限于CLS，并且活性物质和污染物通过膜的交叉速率足够低，则该方法也可以应用于完整细胞。作为工作假设，可以合理地假设单个活性物质的分解速率不会改变，并且切换到全细胞配置的主要效果将是交叉的累加效果。但是，该假设可能会受到各种影响的挑战，包括细胞内电位梯度的差异以及组成成分之间的化学特异性相互作用。图1.（a）体积不平衡的组成对称电池的示意图，在不同体积的储层中，相同的电解质在50％SOC下的浓度相同。活性物质的交换受到抑制，测得的容量衰减可能归因于容量限制侧（CLS）的状态。（b）当电流密度降至1 mA / cm以下时，会发生带极性转换的恒电位循环 这个假设可能会受到各种影响的挑战，包括细胞内电位梯度的差异以及组成成分之间的化学特异性相互作用。图1.（a）体积不平衡的组成对称电池的示意图，在不同体积的储层中，相同的电解质在50％SOC下的浓度相同。活性物质的交换受到抑制，测得的容量衰减可能归因于容量限制侧（CLS）的状态。（b）当电流密度降至1 mA / cm以下时，会发生带极性转换的恒电位循环 这个假设可能受到各种影响的挑战，包括细胞内电位梯度的差异以及组成成分之间的化学特异性相互作用。图1.（a）体积不平衡的组成对称电池的示意图，在不同体积的储层中，相同的电解质在50％SOC下的浓度相同。活性物质的交换受到抑制，测得的容量衰减可能归因于容量限制侧（CLS）的状态。（b）当电流密度降至1 mA / cm以下时，会发生带极性转换的恒电位循环 （a）体积不平衡的组成对称电池的示意图，在不同体积的储层中，相同的电解质在50％SOC下的浓度相同。活性物质的交换受到抑制，测得的容量衰减可能归因于容量限制侧（CLS）的状态。（b）当电流密度降至1 mA / cm以下时，会发生带极性转换的恒电位循环 （a）体积不平衡的组成对称电池的示意图，在不同体积的储层中，相同的电解质在50％SOC下的浓度相同。活性物质的交换受到抑制，测得的容量衰减可能归因于容量限制侧（CLS）的状态。（b）当电流密度降至1 mA / cm以下时，会发生带极性转换的恒电位循环²。从参考文献（26）。版权所有2018。自行车方法。充电步骤和后续放电步骤结束时的恒电位保持对于准确的容量测量至关重要。这可以通过如图1b所示的纯恒电位循环来实现，保持电压直到电流密度下降到非常接近经验确定的背景值（通常为1 mA / cm ²量级）为止。。恒电流循环是充电-放电循环中最常见的方法，它很容易受到电池运行过程中随时间变化而产生的伪影的影响，并且可能掩盖低容量衰减率，如图2所示。尽管这两种技术都需要明智地选择电压极限，但通常会确定通过伏安法和传质分析的结合，以恒定电流运行时，电池的内阻会发生漂移，例如，昼夜温度波动，湿润部件的老化，甚至低SOC时的脉动流也会产生明显的容量褪色。另外，在每个半周期结束时保持恒电流的恒电流循环是消除这些伪影并隔离真实容量衰减率的有效方法（图2）。如果出于任何原因ñ在第三个循环之后，可能保持充电和放电半循环，并且仅将这些循环的放电容量用于评估容量衰减率。如图3所示，在各个SOC处包含停顿可以评估与SOC相关的暂时衰落率，如图3所示。原则上，周期周期的系统变化可以从时间占优的状态中唯一分解出以周期为主导的衰落率褪色率 但是，如图3所示，瞬时衰落速率也可能取决于剩余容量。图2. 2,6-二羟基蒽醌（DHAQ）的体积不平衡组成对称细胞周期的半对数图，显示了纯静电流循环的总电池电阻和电流密度调整后视在容量的变化，并且在恒电位循环和恒电流循环中不存在此类假象，并且在每个半周期结束时均保持了电位。在后两种情况下，通过施加±200 mV的电压来执行循环，并在电流密度降至1 mA / cm时切换²。恒电流条件从10 mA / cm ^2开始并增加到20 mA / cm ²在2.13天之后。垂直箭头指示与电池串联添加和移除电阻器（130mΩ，相当于欧姆电阻增加25％）的时间。当使用严格的恒电流条件时，表观容量的跳跃表明表观循环容量对细胞抗性的依赖性。从参考文献（26）。版权所有2018。图3. 50％SOC 0.1 M DHAQ在1 M KOH中的体积不平衡组成对称的电池循环的半对数图，并在不同电荷状态下出现循环暂停。由每个循环段的斜率得出的瞬时衰减率用黑色文字表示。蓝色文本表示在各种SOC保持期间的衰减率。每20个周期沿上轴用刻度线表示。摘自参考文献（26）。版权所有2018。数据解释。只有在容量衰减到可测量的值以下，该值可低于本体电解实验中最初测量的容量后，才能开始定量解释容量衰减率的测量结果。在此之前，目前尚不清楚是否通过招募出于某种原因在最初几个周期内未获得利用的物种来维持其表观能力。例如，存在残留的溶解氧₂在循环实验开始时，电解液中的有机物可能会通过化学氧化还原的负极电解液而干扰初始循环。如果怀疑在循环试验过程中NCLS可能已达到容量极限，则在循环结束时应将NCLS更换为新鲜的电解质，并施加一些额外的完全充电-放电循环以确认CLS的最终容量。如果骑行方案涉及受限的SOC摆动，则应通过比较受限的SOC方案之前和之后的整个充放电循环来评估其容量衰减率：否则，不清楚是否通过招募物种来维持表观容量在受限的SOC摆动方案期间无法访问。如果容量衰减率太低而无法自信地进行测量，无论是隔离电解质还是工作电池中的电解质，都可以通过提高温度来加速反应。例如，将氧化形式和还原形式（如参考文献（8）中的支持信息中所述，可能需要很好地防止空气接触）在高温下保持数天，尽管很繁琐。执行广泛的化学分析（例如质谱，NMR和循环伏安法）。但是，由于主要机制可能与温度有关，因此应谨慎。另外，必须区分氧化还原活性和非活性分解产物。此外，隔离中的活性物质寿命可能不同于与操作池中的湿润材料接触的寿命。隔离的电解液或工作电池中的电解液。例如，将氧化形式和还原形式（如参考文献（8）中的支持信息中所述，可能需要很好地防止空气接触）在高温下保持数天，尽管很繁琐。执行广泛的化学分析（例如质谱，NMR和循环伏安法）。但是，由于主要机制可能与温度有关，因此应谨慎。另外，必须区分氧化还原活性和非活性分解产物。此外，隔离中的活性物质寿命可能不同于与操作池中的湿润材料接触的寿命。隔离的电解液或工作电池中的电解液。例如，将氧化形式和还原形式（如参考文献（8）中的支持信息中所述，可能需要很好地防止空气接触）在高温下保持数天，尽管很繁琐。执行广泛的化学分析（例如质谱，NMR和循环伏安法）。但是，由于主要机制可能与温度有关，因此应谨慎。另外，必须区分氧化还原活性和非活性分解产物。此外，隔离中的活性物质寿命可能不同于与操作池中的湿润材料接触的寿命。以便将氧化形式和还原形式（如参考文献（8）中的支持信息中所述，应很好地保护空气免受氧化）在高温下保持多天，并进行广泛的化学分析（例如质谱分析， NMR和循环伏安法）。但是，由于主要机制可能与温度有关，因此应谨慎。另外，必须区分氧化还原活性和非活性分解产物。此外，隔离中的活性物质寿命可能不同于与操作池中的湿润材料接触的寿命。以便将氧化形式和还原形式（如参考文献（8）中的支持信息中所述，应很好地保护空气免受氧化）在高温下保持多天，并进行广泛的化学分析（例如质谱分析， NMR和循环伏安法）。但是，由于主要机制可能与温度有关，因此应谨慎。另外，必须区分氧化还原活性和非活性分解产物。此外，隔离中的活性物质寿命可能不同于与操作池中的湿润材料接触的寿命。应谨慎，因为主要机制可能与温度有关；另外，必须区分氧化还原活性和非活性分解产物。此外，隔离中的活性物质寿命可能不同于与操作池中的湿润材料接触的寿命。应谨慎，因为主要机制可能与温度有关；另外，必须区分氧化还原活性和非活性分解产物。此外，隔离中的活性物质寿命可能不同于与操作池中的湿润材料接触的寿命。数据报告。建议同时报告每天和每个周期的容量衰减率。如果通常观察到一定程度的容量损失，但不是唯一分解为以周期和时间表示的速率，则容量衰减率可以报告为“每天X％或每个周期Y％”。应该报告容量衰减率而不是容量保留率，因为在小时，每天和每年的衰减率之间进行转换的算法比在相应容量保留率之间进行转换的算法更简单。重要的是要报告有关实验方法的足够信息，包括电解质量，以使知识渊博的同事能够重现实验结果。成本-终生权衡。在某些情况下，对于有机化学而言，与现有的钒化学物质相比，其成本可能不会更低，因此分子的寿命不一定是无限的。的确，可以将使用廉价的有机物代替钒的前期资本成本节省与一系列未来重置成本的现值进行比较，从而确认货币的时间价值。权衡是通过重置成本比率来量化的，重置成本比率定义为年度重置成本除以前期资本成本节省。（2）重置成本比率的损益平衡值取决于折现的利率和项目寿命，如图4所示。当重置成本比率的实际值小于此收支平衡值时，有机物的成本可能低于钒。例如，如果有机物占钒每千瓦时成本的40％，但每年损失9％，则替代成本比率为（0.09）（0.4）/（1.0 – 0.4）= 0.06，而20年的项目收支平衡利率为1.8％，较高的利率有利于有机债券。如果人们期望通过定期维护电解液来实现钒的无限期项目，那么在与以前相同的假设下，将以6.0％的利率实现收支平衡。这种简单的比较假设当有机或钒替代物出现时，翻新成本为零，但是可以通过增加翻新成本来推广该模型。图4.假定项目寿命如所示，重置成本比率的收支平衡值与折现利率的关系。改编自参考文献（2）。版权所有2018。则重置成本比率为（0.09）（0.4）/（1.0 – 0.4）= 0.06，并且20年期项目的收支平衡为1.8％，较高的利率有利于有机成本。如果人们期望通过定期维护电解液来实现钒的无限期项目，那么在与以前相同的假设下，将以6.0％的利率实现收支平衡。这种简单的比较假设当有机或钒替代物出现时，翻新成本为零，但是可以通过增加翻新成本来推广该模型。图4.假设项目寿命如所示，重置成本比率的收支平衡值与折现率的关系。改编自参考文献（2）。版权所有2018。则重置成本比率为（0.09）（0.4）/（1.0 – 0.4）= 0.06，而20年期项目的收支平衡为1.8％，较高的利率有利于有机债券。如果人们期望通过定期维护电解液来实现钒的无限期项目，那么在与以前相同的假设下，将以6.0％的利率实现收支平衡。这种简单的比较假设当有机或钒替代物出现时，翻新成本为零，但是可以通过增加翻新成本来推广该模型。图4.假设项目寿命如所示，重置成本比率的收支平衡值与折现率的关系。改编自参考文献（2）。版权所有2018。高利率有利于有机。如果人们期望通过定期维护电解液来实现钒的无限期项目，那么在与以前相同的假设下，将以6.0％的利率实现收支平衡。这种简单的比较假设当有机或钒替代物出现时，翻新成本为零，但是可以通过增加翻新成本来推广该模型。图4.假设项目寿命如所示，重置成本比率的收支平衡值与折现率的关系。改编自参考文献（2）。版权所有2018。高利率有利于有机。如果人们期望通过定期维护电解液来实现钒的无限期项目，那么在与以前相同的假设下，将以6.0％的利率实现收支平衡。这种简单的比较假设当有机或钒替代物出现时，翻新成本为零，但是可以通过增加翻新成本来推广该模型。图4.假定项目寿命如所示，重置成本比率的收支平衡值与折现利率的关系。改编自参考文献（2）。版权所有2018。这种简单的比较假设当有机或钒替代物出现时，翻新成本为零，但是可以通过增加翻新成本来推广该模型。图4.假设项目寿命如所示，重置成本比率的收支平衡值与折现率的关系。改编自参考文献（2）。版权所有2018。这种简单的比较假设当有机或钒替代物出现时，翻新成本为零，但是可以通过增加翻新成本来推广该模型。图4.假定项目寿命如所示，重置成本比率的收支平衡值与折现利率的关系。改编自参考文献（2）。版权所有2018。有限寿命的电解质设计空间。有限的电解质寿命和必要的替代品的影响可以整合到现有的资本成本模型中，以说明一系列未来化学物质替代成本的现值，并概述化学物质成本的允许范围，褪色率和折现率及其依赖性（28,29）为简洁起见，支持信息中详细介绍了数学处理，简化假设和基准模型输入。在这里，我们说明了定期更换电解质对AORFBs经济可行性的影响，该影响是容量衰减率，折现率，放电持续时间和氧化还原化学性质（例如，电池电势，化学成本和电池电阻）的函数（图5）。如图5a所示，只要氧化还原化学物质具有高电池电压和低电解质成本的适当组合，就可以提供有限寿命的有机物窗口。后者可以通过平衡廉价的化学品，低褪色率和优惠的贴现利率来实现（图5b）。可视化资本成本对总活动成本的依赖性和敏感性，作为电池电势，电池电阻和持续时间的函数，还可以洞悉反应堆性能和成本在定义具有成本效益的应用空间方面的相对重要性（图5c，d）。图5.包含有限寿命电解质的净现值的通用RFB的资本成本模型。（a）随着放电时间的延长，功率更高的反应堆和廉价电解质的使用，资本成本降低。在长时间（10–100小时）中，系统成本开始渐近地接近储层中电解质的成本。作为比较，提供了一系列近期钒历史价格中钒RFB的成本。（b）总有效成本与电解质成本成正比，是化学成本，褪色率和贴现率的函数。各种放电持续时间和（c）电池电势的资本成本与总有效成本；（d）特定区域电阻（ASR）。以较低的总活动成本，电池的资本成本显示出对电池电势，放电持续时间和电池电阻的更大敏感性。在高昂的总活跃成本下，电解质主导了资本成本，除了电池电压会影响功率和能源成本外，其他参数之间几乎没有区别。对于面板a–c中显示的数据，电池ASR为1.0Ω·cm假定为²，并且对于面板d，假定为1.0V的电池电压。有关模型输入，公式化和数据可视化的更多详细信息，请参见支持信息。概要。用于RFB的电解质材料的成本是具有高E / P比的系统的主要因素。对于其中溶剂和盐通常廉价的水性体系，活性材料的成本是最令人关注的。由富含地球的元素制成的有机活性材料在合成后可以（但不一定）是低成本的。但是，它们的有限寿命会增加定期更换成本，从而增加成本，从而使技术经济评估变得复杂。目前，我们已经介绍了有关如何评估有机活性材料寿命的最佳方法。我们已经展示了如何将寿命纳入每千瓦时能量存储容量的系统成本估算中。总活动成本与活性材料每千瓦时的资本成本成正比，并且是年度重置比率和将未来费用折现至当前的利率的函数。现有钒RFB的成本可作为比较的便利。这些结果开始为具有成本竞争力的AORFB定义了一个设计空间，该空间需要稳定，经济的具有足够不同氧化还原电位的材料。在文献中报道的众多且不断扩展的有机氧化还原对中，（35）可以鉴定出一些具有高电池电位（≥1 V）和稳定性（每年约5％或更低）的氧化还原化学物质（8），尽管据作者所知，没有一个经过了一年的测试。在平行下，Yang等人最近的初步工艺经济评估。和Dieterich等。已显示出低成本材料的潜在途径（每侧约35美元/ kAh或更低），（2,3），但这些材料通常与更高的褪色率相关。随着该领域的发展，需要进一步的工作来阐明材料成本，稳定性和易于回收之间的科学和经济折衷。到最后一点，最近的工作表明了原位再生（36）最终可以用作容量恢复机制，类似于钒RFB中的电解质再平衡，从而进一步改善了具有中等衰减率的潜在廉价材料的前景。最终，作者希望这一观点为研究界提供指导，并激发新的研究方向，从而进一步促进AORFB的科学和工程学。可从https://pubs.acs.org/doi/10.1021/acsenergylett.0c00140免费获得支持信息。

• 关于这项工作中进行的技术经济分析的数学公式，模型输入和简化假设（PDF）

关于这项工作中进行的技术经济分析的数学公式，模型输入和简化假设（PDF）MJA：概念化，方法论，书面起草，监督，资金筹集；FRB：方法论，起草原件，资源，监督，资金获取；KER：方法论，软件，调查，可视化。本观点表达的观点仅为作者的观点，不一定是ACS的观点。作者宣称没有竞争性的经济利益。无需订阅ACS Web版本即可获得电子支持信息文件。美国化学学会在任何可版权保护的支持信息中拥有版权权益。ACS网站上提供的文件只能下载供个人使用。未经美国化学学会许可，不得以其他方式允许用户以机器可读形式或任何其他形式全部或部分复制，重新发布，重新分发或出售ACS网站上的任何支持信息。为了获得复制，重新发布和重新分发此材料的许可，请求者必须通过RightsLink许可系统处理自己的请求。有关如何使用RightsLink权限系统的信息，请访问http://pubs.acs.org/page/copyright/permissions.html。哈佛大学的工作得到了美国能源部授予的DE-AC05-76RL01830（通过PNNL分包合同428977），ARPA-E授予的DE-AR0000767（由丹麦创新基金通过大解决方案项目“ ORBATS”文件nr）的支持。7046-00018B，并由马萨诸塞州清洁能源技术中心提供。麻省理工学院的工作得到了能源存储研究联合中心（JCESR）的支持，该中心是由美国能源部科学技术办公室，基础能源科学（DE-AC02-06CH11357）资助的能源创新中心。作者感谢Eric Fell，Martin Jin，Yan Jing，Emily Kerr，Daniel Pollack，Tang Jiang，Andrew Wong和Min Min（哈佛）以及Michael Orella（MIT）的有益评论。本文引用了其他36个出版物。