No abstract

Redox flow batteries (RFBs) are propitious stationary energy storage technologies with exceptional scalability and flexibility to improve the stability, efficiency, and sustainability of our power grid. The redox-active materials are the key component for RFBs with which to achieve high energy density and good cyclability. Traditional inorganic-based materials encounter critical technical and economic limitations such as low solubility, inferior electrochemical activity, and high cost. Redox-active organic materials (ROMs) are promising alternative "green" candidates to push the boundaries of energy storage because of the significant advantages of molecular diversity, structural tailorability, and natural abundance. Here, the recent development of a variety of ROMs and associated battery designs in both aqueous and nonaqueous electrolytes are reviewed. The critical challenges and potential research opportunities for developing practically relevant organic flow batteries are discussed.

Aqueous redox flow batteries with organic active materials offer an environmentally benign, tunable, and safe route to large-scale energy storage. Development has been limited to a small palette of organics that are aqueous soluble and tend to display the necessary redox reversibility within the water stability window. We show how molecular engineering of fluorenone enables the alcohol electro-oxidation needed for reversible ketone hydrogenation and dehydrogenation at room temperature without the use of a catalyst. Flow batteries based on these fluorenone derivative anolytes operate efficiently and exhibit stable long-term cycling at ambient and mildly increased temperatures in a nondemanding environment. These results expand the palette to include reversible ketone to alcohol conversion but also suggest the potential for identifying other atypical organic redox couple candidates.

No abstract

impact the performances of RFBs. Traditionally, aqueous RFBs based on inorganic materials with an emphasis on metal species have been extensively developed, but suffer substantially from limitations, such as low cell voltage, inferior energy density, and high cost. [1] To push the boundaries of grid storage, nonaqueous RFBs are being investigated to pursue high energy density, enabled by the wide voltage windows of nonaqueous electrolytes that lead to higher cell voltages and more choices of materials candidates. [2] A variety of redox-active materials and cell designs have been developed for nonaqueous RFBs including metal coordination complexes, [3] redox-active polymers, [4] semisolid suspensions, [5] Li metal RFBs, [6] and redox targeting RFBs. [7] Notably, redox-active organic materials (ROMs) have gained significant momentum for nonaqueous RFBs. [8] This is mainly ascribed to the great potential of ROMs in expanding the library of material candidates and tailoring the energy density and cyclability of RFBs via molecular engineering. Currently, considerable progress has been achieved in the discovery and prototyping of promising ROMs for nonaqueous RFBs. [9] However, one of the major challenges for ROM-based nonaqueous RFBs is low energy density. Adopting ROMs that allow multiple electron transfers for RFBs is an effective strategy to boost the energy density. A few ROMs, including quinones, N-alkylated pyridiniums, and alloxazines, have been reported to show 2e − transfers in nonaqueous or aqueous RFBs. [10] In this communication, we demonstrate a new two-electron nonaqueous organic RFB based on phenothiazine (PT) catholyte and anthraquinone (AQ) anolyte materials. PTs have been demonstrated to have stable 1e − reactions for nonaqueous RFBs and lithium ion battery overcharge protection applications. [9c,11] Our studies indicate that PT also features 2e − transfers in nonaqueous electrolytes. When coupled with AQ derivatives, a proof-of-concept 2e − PT/AQ RFB was demonstrated with decent cyclability. PT is oxidized by stepwise losing 2e − to form the PT 2+ dication during charging while AQ is reduced by stepwise acquiring 2e − to form the AQ 2− dianion. Reverse processes occur during discharging.A series of structurally tailored AQ and PT derivatives were prepared using the synthetic approaches illustrated in Scheme 1 and the Supporting Information. AQs were synthesized in one step from the corresponding hydroxyanthraquinones and 2-bromoethyl methyl ether with high yields under the treatment of Cs 2 CO 3 in N,N-dimethylformamide (DMF).

Redox-active organic materials (ROMs) are becoming increasingly attractive for use in redox flow batteries as promising alternatives to traditional inorganic counterparts. However, the reported ROMs are often accompanied by challenges, including poor solubility and stability. Herein, we demonstrate that the commonly used diquat herbicides, with solubilities of >2 M in aqueous electrolytes, can be used as stable anolyte materials in organic flow batteries. When coupled with a ferrocene-derived catholyte, the flow cells with the diquat anolyte demonstrate long galvanic cycling with high capacity retention. Notably, the mechanistic underpinnings of this remarkable stability are attributed to the improved π-conjugation that originated from the near-planar molecular conformations of the spatially constrained 2,2′-bipyridyl rings, suggesting a viable structural engineering strategy for designing stable organic materials.

Aqueous organic redox flow batteries are a promising technology for large-scale energy storage. The stability of the redox active organic molecules is increasingly being recognized as one of the major...

Determining the aqueous solubility of molecules is a vital step in many pharmaceutical, environmental, and energy storage applications. Despite efforts made over decades, there are still challenges associated with developing a solubility prediction model with satisfactory accuracy for many of these applications. The goals of this study are to assess current deep learning methods for solubility prediction, develop a general model capable of predicting the solubility of a broad range of organic molecules, and to understand the impact of data properties, molecular representation, and modeling architecture on predictive performance. Using the largest currently available solubility data set, we implement deep learning-based models to predict solubility from the molecular structure and explore several different molecular representations including molecular descriptors, simplified molecular-input line-entry system strings, molecular graphs, and three-dimensional atomic coordinates using four different neural network architectures—fully connected neural networks, recurrent neural networks, graph neural networks (GNNs), and SchNet. We find that models using molecular descriptors achieve the best performance, with GNN models also achieving good performance. We perform extensive error analysis to understand the molecular properties that influence model performance, perform feature analysis to understand which information about the molecular structure is most valuable for prediction, and perform a transfer learning and data size study to understand the impact of data availability on model performance.