A High‐Voltage Membrane‐Free Li–Organic Hybrid Flow Battery Using Eutectic Lithium Chemistry Authors: Xiao Wang, Bindu Dahal, Yuhang Tang, Jack J McGrath, Rabin Siwakoti, Feng Wang, Jun Cheng, Jianbing “Jimmy” Jiang Published: 2026 ABSTRACT ABSTRACTMembrane‐free redox flow batteries offer simplified design and reduced cost, but their electrochemical performance remains limited. Here, a high‐voltage nonaqueous membrane‐free Li–organic hybrid flow battery was constructed using a biphasic electrolyte. The bottom‐phase anolyte is a deep eutectic solvent formed by lithium hexafluorophosphate and 2,2,2‐trifluoroacetamide, which provides a wide electrochemical stability window of 4.81 V, low viscosity of 13.8 cP, high ionic conductivity of 8.59 mS·cm−1, and stable compatibility with lithium metal. The top‐phase catholyte is dichloromethane containing a long‐alkyl‐chain phenothiazine derivative, which exhibits excellent redox reversibility and strict confinement in the upper phase. The biphasic system achieved an open‐circuit voltage of approximately 3.6 V and stable operation at 0.5 M concentration, retaining 85.6% of theoretical capacity with a Coulombic efficiency of 97.6% after 100 cycles. In flow configuration, the 0.5 M battery maintained over 92% capacity after 10 days of continuous cycling with minimal self‐discharge. A preliminary cost estimation yielded $134.8 kWh−1, which is lower than vanadium redox flow batteries, and both electrolyte phases displayed strong flame resistance. This study provides a viable design strategy to enable membrane‐free nonaqueous Li–organic hybrid flow battery by leveraging eutectic lithium chemistry and immiscible solvent interfaces.Keywords: biphase, deep eutectic solvent, Li–organic hybrid flow battery, lithium anode, membrane‐free Introduction Redox flow batteries (RFBs) have garnered significant attention for grid‐scale energy storage due to their scalable architecture, long cycle life, and decoupled energy and power design [1, 2, 3, 4, 5, 6, 7, 8, 9]. These systems are particularly well suited for balancing the intermittency of renewable energy sources, such as wind and solar, by enabling peak shaving and load leveling [10, 11, 12, 13, 14, 15]. Conventional RFBs typically employ ion‐exchange membranes to separate the anolyte and catholyte, ensuring selective ion transport while preventing the crossover of redox‐active species [16, 17, 18, 19]. However, membranes substantially increase system cost, add internal resistance, and pose challenges in long‐term chemical compatibility, especially under nonaqueous conditions where membrane swelling, degradation, or incompatibility with organic solvents may occur [20]. To address these limitations, membrane‐free redox flow batteries have emerged as a promising alternative that eliminates the membrane altogether [21, 22, 23, 24, 25, 26, 27, 28]. In these systems, phase immiscibility serves as the physical barrier between redox‐active electrolytes. This approach simplifies cell architecture, reduces manufacturing cost, and removes the dependency on expensive and failure‐prone membrane materials. Achieving effective phase separation and long‐term interfacial stability requires careful design of solvent systems that combine immiscibility, interfacial robustness, selective redox confinement, and sufficient ionic conductivity [29]. Several types of membrane‐free biphasic systems have been explored to date, including aqueous–nonaqueous, aqueous–aqueous, and nonaqueous–nonaqueous combinations. Each of these architectures offers unique benefits and faces distinct challenges. Aqueous–nonaqueous systems benefit from clear phase separation driven by intrinsic polarity differences and typically exhibit low viscosity and fast ion transport. However, their cell voltages are restricted to ∼1.2–2.0 V due to the narrow electrochemical window of water, and interfacial compatibility issues can arise from solubility mismatches or parasitic reactions at the interface [30, 31, 32]. Aqueous–aqueous systems, particularly those based on salt‐induced biphasic separation (such as LiTFSI/LiCl aqueous mixtures) [33], have attracted attention due to their inherent safety, nonflammability, and use of water‐compatible materials. However, they are fundamentally constrained by the limited electrochemical stability window of water, and the high concentration of salts required to maintain phase separation, which increases cost and viscosity [23, 33]. Nonaqueous‐nonaqueous biphasic systems have also been developed to access broader electrochemical windows (> 3.0 V) and accommodate a wider range of organic redox‐active species. These systems offer improved solubility and voltage tunability compared to aqueous counterparts. However, their implementation often relies on high‐cost and viscous solvents such as ionic liquids or fluorinated ethers [29, 34]. While these solvents offer excellent electrochemical stability, they suffer from drawbacks including high viscosity, complex synthesis, and limited availability, all of which raise material and processing costs. As a result, despite their electrochemical advantages, the scalability and cost‐effectiveness of fully nonaqueous membrane‐free systems remain limited. In this context, deep eutectic solvents (DESs) offer a highly attractive alternative. Formed by combining lithium salts and hydrogen bond donors, DESs exhibit physicochemical properties similar to ionic liquids, including broad electrochemical windows (often exceeding 4.0 V), negligible vapor pressure, and intrinsic ionic conductivity, while offering significantly lower cost and greater compositional flexibility [35]. Recent studies have demonstrated DESs based on lithium salts and amide‐type hydrogen bond donors with conductivities exceeding 10 mS·cm−1, enabling promising performance in lithium‐based electrochemical systems [36, 37]. Furthermore, their relatively high densities facilitate stable biphasic stratification when paired with low‐density halogenated solvents. These characteristics render DESs ideal candidates for use as bottom‐phase electrolytes in membrane‐free biphasic flow batteries, particularly when cost, interfacial stability, and ionic mobility are key design considerations. In this work, we report a high‐voltage membrane‐free nonaqueous Li–organic hybrid flow battery based on a biphasic electrolyte system that combines low cost, low viscosity, and high electrochemical performance. The bottom‐phase anolyte is a deep eutectic solvent composed of lithium hexafluorophosphate and 2,2,2‐trifluoroacetamide (TFAD), with tunable composition (4:1 to 5:1) to ensure liquid‐state stability across different states of charge. This eutectic solvent exhibits a remarkably low viscosity of 13.8 cP and significantly reduced cost compared to conventional ionic liquids. The top‐phase catholyte consists of a long‐alkyl‐chain‐substituted phenothiazine (C8‐PTZ) dissolved in saturated tetrabutylammonium hexafluorophosphate (TBAPF6) in dichloromethane (DCM). The strong immiscibility between the two liquid phases forms a stable, clear interface that effectively confines redox reactions and eliminates the need for a membrane. The resulting system achieves an open‐circuit voltage of ∼3.6 V, Coulombic efficiency above 97%, and stable operation over 100 cycles under both static and flow conditions. This design highlights a practical, cost‐effective pathway for building membrane‐free, high‐voltage Li–organic hybrid flow batteries using eutectic lithium electrolytes and immiscible solvent interfaces. Results and Discussion Designing of Electrolyte Systems As illustrated in Figure 1A, the membrane‐free biphasic Li–organic hybrid flow battery was constructed using a TFAD/LiPF6 eutectic solvent as the anolyte (bottom phase) and a phenothiazine derivative, C8‐PTZ, in DCM/TBAPF6 as the catholyte (top phase). To rationally design such a biphasic system, a series of DESs were first formulated and screened to ensure stable phase separation and electrochemical compatibility with lithium metal. These DESs were prepared by combining different lithium salts: LiTFSI, LiPF6, and LiClO4, with hydrogen bond donors including urea, acetamide, and TFAD, as shown in Figure S1. The selection criteria for a viable DES included four key properties: (i) a wide electrochemical stability window to allow safe operation with lithium metal; (ii) low viscosity for reduced ohmic losses and ease of pumping; (iii) high ionic conductivity to ensure sufficient current flow; and (iv) liquid‐phase stability at room temperature for flow operation. (A) Schematic of the membrane‐free biphasic Li–organic hybrid flow battery using TFAD/LiPF6 (5:1) as the anolyte (bottom phase) and DCM/TBAPF6 with C8‐PTZ as the catholyte (top phase). (B) Voltage profile of a symmetric Li||Li cell cycled in TFAD/LiPF6 (5:1) DES at a current density of 0.8 mA cm−2. (C) Raman spectra of neat TFAD and TFAD/LiPF6 (5:1). (D) FT‐IR spectra of neat TFAD and TFAD/LiPF6 (5:1) at a scan range of 2500 to 3800 cm−1. Electrochemical evaluations of different DESs (Figures S2–S6) showed that the electrochemical stability window varied significantly depending on both the hydrogen bonding donator and the lithium salt. Among the various combinations, many DESs demonstrated broad electrochemical stability windows. Systems such as acetamide/LiPF6 (4:1), acetamide/LiTFSI (4:1), and urea/LiTFSI (3:1) exhibited anodic stability exceeding 4.4 V (Figures S4–S6). Notably, the TFAD/LiPF6 eutectic at a 5:1 molar ratio not only achieved an electrochemical window up to 4.81 V (Figure S2) but also displayed distinct lithium plating and stripping features, indicating excellent compatibility with lithium redox processes. The physical properties of the DESs, including viscosity, density, and phase state, were systematically evaluated to guide electrolyte selection (Table S1). Low viscosity is essential for promoting efficient ion transport, reducing ohmic losses, and facilitating flow operation. Among the candidates, TFAD/LiPF6 combination exhibited notably low viscosity of 13.8 cP, significantly lower than acetamide‐based (over 70 cP) or urea‐based (over 1000 cP) systems, highlighting its superior fluidity. In addition to viscosity, density plays a critical role in determining phase arrangement and stability in biphasic systems. DESs generally exhibited higher densities compared to conventional organic solvents, which is advantageous for forming a stable lower‐phase electrolyte. Notably, acetamide/LiClO4 (4:1) reached a density as high as 2.075 g·cm−3, and most of the DES still exceeded 1.2 g·cm−3. This overall high‐density profile ensures that the DES remains well separated from common lighter electrolyte solvents such as DCM, MeCN, and carbonate, which facilitates reliable phase stratification. These physical characteristics, particularly the combination of low viscosity and relatively high density, highlight the structural advantages of DESs in constructing robust biphasic battery systems without the need for a membrane. A series of electrochemical impedance spectroscopy (EIS) measurements were conducted on different electrolyte systems to evaluate their ionic conductivity (Figure S7 and Table S1). Except for the urea‐based DESs, which exhibited excessively high viscosity and poor conductivity, all other systems demonstrated ionic conductivities on the order of 1 mS·cm−1. Among them, the TFAD/LiPF6 eutectic displayed one of the lowest impedance responses and achieved the highest conductivity of 8.59 mS·cm−1. The chemical compatibility of the DESs with lithium metal was evaluated through direct contact experiments after 24 h. Based on their favorable electrochemical stability, selected formulations were tested for lithium tolerance. As shown in Figure S8, lithium metal retained its metallic luster and showed no apparent signs of surface corrosion or passivation after immersion in TFAD/LiPF6. In contrast, visible degradation was observed when lithium was exposed to other DES candidates, suggesting that TFAD‐based DESs offer a chemically inert environment for lithium metal. The improved stability may be attributed to the lower hydrogen content and higher fluorine exposure to Li metal, as fluorinated environments are known to exhibit reduced reactivity toward lithium in the absence of electrochemical processes [38, 39, 40]. This compatibility with Li metal is further supported by the symmetric Li||Li cycling results in Figure 1B, which demonstrates the smooth and stable lithium plating/stripping behavior over extended cycling. Considering the combined advantages of wide electrochemical stability, low viscosity, high ionic conductivity, appropriate density, and good compatibility with lithium metal under the tested conditions, the TFAD/LiPF6 eutectic system was selected. Together, the immersion and symmetric‐cell results support good compatibility between the TFAD/LiPF6 electrolyte and lithium metal under the tested conditions. Nevertheless, these measurements provide a preliminary assessment rather than a comprehensive characterization of the lithium–electrolyte interface, and more detailed studies of lithium deposition efficiency and interfacial chemistry would be valuable in future work. To gain molecular‐level insight into the interactions within the selected TFAD/LiPF6 (5:1) eutectic system, we conducted spectroscopic analyses using IR, Raman, and NMR analyses. Raman spectroscopy was employed to investigate the molecular interactions within the TFAD/LiPF6 eutectic system. As shown in Figure 1C, the C = O stretching vibration of neat TFAD appears at approximately 1670 cm−1. Upon mixing with LiPF6 at a 5:1 molar ratio, this band exhibits a slight redshift, indicative of weakened C = O bond strength due to coordination with Li+ ions or involvement in hydrogen bonding. In addition, the N–H stretching band observed near 3300 cm−1 in neat TFAD is significantly attenuated in the eutectic mixture, suggesting that the N–H moiety is engaged in hydrogen bonding or coordination, leading to loss of Raman activity due to increased vibrational asymmetry. In contrast, the characteristic PF6 − symmetric stretching mode (∼740 cm−1) and bending modes (< 600 cm−1) remain largely unchanged in the mixture, indicating that PF6 − anions retain their structural integrity and do not participate strongly in coordination with TFAD. This observation supports the conclusion that the weakly coordinating nature of PF6 − contributes to a more dissociated ionic environment and lower viscosity. Moreover, the δ(CF3) bending modes remain unaffected, confirming that the trifluoromethyl group plays a primarily electronic and steric role rather than being chemically interactive. Together, these spectral features provide direct evidence of moderate hydrogen bonding and minimal ion aggregation, which underlie the favorable flow properties of the eutectic system. In the IR spectra of neat TFAD, a prominent C = O stretching vibration signal appears at 1670 cm−1, accompanied by sharp N–H stretching bands between 3300–3400 cm−1, indicating a well‐defined hydrogen‐bonding network primarily involving N–H···O = C interactions (Figures 1D and S10). Upon addition of LiPF6, the C = O band exhibits a distinct redshift and broadening in both IR and Raman spectra, strongly suggesting coordination of the carbonyl oxygen to Li+ ions. This coordination weakens the double bond character of the C = O bond, lowering its vibrational frequency and producing heterogeneous local environments. Simultaneously, the N–H stretching region undergoes a global redshift, peak broadening, and increased IR absorbance, while its Raman signal diminishes significantly. These changes indicate disruption of the original N–H···O = C hydrogen bonds and the formation of new, less symmetric hydrogen bonding or coordination involving PF6 − or reorganized TFAD molecules. The broadening and intensity loss in this region imply vibrational dephasing and increased asymmetry, consistent with a more disordered environment in the eutectic liquid. In the low‐wavenumber Raman region, the δ(CF3) bending vibration of TFAD at ∼412 cm−1 is significantly suppressed in the mixture, indicating structural disordering and vibrational incoherence due to eutectic melting. In addition, the C–N stretching band at ∼1113 cm−1 becomes broader and weaker upon mixing, reflecting the breakdown of ordered hydrogen bonding and the emergence of diverse coordination geometries around the amide nitrogen. In contrast, PF6 − anions retain their structural identity, as evidenced by the preservation of their symmetric stretching mode (∼740 cm−1) and bending modes (<600 cm−1) in the Raman spectrum. These minimal spectral changes suggest that PF6 − does not actively participate in strong coordination or hydrogen bonding with TFAD, reinforcing its role as a weakly coordinating anion. This helps maintain high ionic dissociation and contributes to the low viscosity observed in the eutectic system. these spectral features reveal a clear transition from an ordered hydrogen‐bonded solid to a dynamic, partially coordinated ionic liquid‐like environment, underpinned by C = O···Li+ coordination and N–H hydrogen bonding reorganization. This transformation accounts for the favorable flowability and ion transport properties of the TFAD/LiPF6 eutectic mixture. In the 1 9F NMR spectra of TFAD/LiPF6 DES, PF6 − consistently displays two equally integrated resonances at ∼−71.5 and ∼−73.5 ppm. Although PF6 − possesses perfect octahedral symmetry, its non‐rigid nature on the NMR timescale allows spin–spin interactions among equivalent fluorine nuclei to generate a symmetric doublet‐like pattern with equal intensity, a phenomenon well documented for other non‐rigid polyatomic anions [41, 42]. At room temperature, only a single averaged spectral response is observed for PF6 −, most likely because different ionic environments (e.g., contact ion pairs (CIP) and solvent‐separated ion pairs (SSIP)) interconvert rapidly, resulting in fast or intermediate exchange that yields weighted‐average signals in NMR, Raman, and IR spectra. To further probe this dynamic equilibrium, variable‐temperature NMR measurements were performed. At 15°C, distinct shoulder features emerged on both PF6 − resonances across all examined compositions (Figure S11), providing evidence for the presence of different ionic sub‐populations in the eutectic system. In addition to maintaining the characteristic doublet splitting, the two PF6 − resonances exhibit a systematic upfield shift as the TFAD:LiPF6 molar ratio increases from 3.5:1 to 5:1 (Figures 2A and S12). This shift reflects a gradual change in the average electronic environment of PF6 −: as more TFAD molecules solvate Li+, the Li+–PF6 − electrostatic association weakens, allowing PF6 − to reside further from the Li+ solvation shell and become more electronically shielded. A separate resonance at ∼−77.63 ppm corresponds to the CF3 group of TFAD, in good agreement with the value obtained from solid‐state NMR (Figures 2B and S13). The invariance of this peak confirms that the CF3 group remains chemically inert during eutectic formation and does not participate in coordination or hydrogen bonding. Notably, the CF3 resonance also exhibits a subtle systematic upfield shift (∼0.05 ppm) as the TFAD:LiPF6 ratio increases, suggesting enhanced electronic shielding of CF3 due to reduced hydrogen‐bonding density and an increased fraction of uncoordinated TFAD molecules at higher ratios. In the high‐field region of the 1H NMR spectra (Figure S14), both neat TFAD and its 4:1 deep eutectic mixture with LiPF6 exhibit two well‐resolved peaks. These two peaks originate from the two amide protons that are magnetically non‐equivalent due to restricted rotation around the amide C–N bond, a phenomenon widely observed in primary amides [43]. Upon coordination with lithium salt to form the 4:1 DES, the chemical shifts of the N–H protons show negligible change compared to neat TFAD in DMSO‐d6, indicating that the amide hydrogens are not significantly involved in coordination with Li+ in DMSO. This suggests that Li+ preferentially interacts with the carbonyl oxygen rather than with the –NH2 protons when a strong donor solvent such as DMSO is present. (A) 19F NMR spectra of TFAD/LiPF6 DESs at different molar ratios (3.5:1 to 5:1) from −74 to −71 ppm. (B) Solid‐state 19F NMR spectra of TFAD. (C) Snapshot from molecular dynamics simulation of LiPF6/TFAD (5:1). (D) Radial distribution functions (RDF) of Li+–TFAD and Li+–PF6 − interactions. (E) Coordination number profiles between Li+ and TFAD atoms (O, N, F). (F) Solvation structure distribution for TFAD‐based deep eutectic solvents with LiPF6, LiClO4, and LiTFSI, as determined from molecular dynamics simulations. (G) Arrhenius plot and its fitting curve of ionic conductivity with a linear fit of ln σ vs. 1/T. The solvation structures of Li+ ions in the three electrolyte systems were systematically analyzed through radial distribution function (RDF) calculations and coordination number analysis. The Li+‐TFAD RDF profiles show distinct first peaks at approximately 2.0 Å (Figure 2C,D), corresponding to Li+ coordination with carbonyl oxygen atoms, followed by a second peak around 2.8 Å attributed to interactions with trifluoromethyl (‐CF3) groups (Figure 2E). The first solvation shell boundary was defined at 3 Å based on the minimum following the primary coordination peaks in the RDF profiles. Analysis of the Li+‐TFAD coordination within this 3 Å cutoff reveals that the interactions are predominantly governed by carbonyl oxygen (C = O) coordination, with minor contributions from CF3 groups in proximity. The first solvation shell coordination numbers, determined from RDF integration, reveal distinct solvation environments: Li(PF6)1.352(TFAD)3.389 for the LiPF6‐TFAD system, Li(ClO4)1.879(TFAD)3.209 for the TFAD/LiClO4 system, and Li(TFSI)2.203(TFAD)2.428 for the TFAD/LiTFSI system (Figure S15). A progressive increase in anion coordination with Li+ is observed across the series (PF6 − < ClO4 − < TFSI−), which can be attributed to the increasing coordinating ability and size of the anions. This enhanced anion‐cation interaction correspondingly reduces the available coordination sites for TFAD molecules, resulting in the observed decrease in TFAD solvation numbers. The coordination number increases beyond 3 Å primarily arises from secondary interactions with trifluoromethyl (‐CF3) groups of TFAD molecules in the outer solvation sphere. The weaker coordination of PF6 − anions with Li+ facilitates greater solvent participation in the solvation shell, promoting the formation of solvent‐mediated ion pairs. In contrast, the stronger coordinating anions (ClO4 − and TFSI−) tend to form more intimate ion pairs and aggregates due to their enhanced electrostatic interactions with Li+ ions, thereby displacing solvent molecules from the primary coordination sphere. Analysis of ion association patterns reveals significant variations in solvation structures among the three systems, with the TFAD/LiPF6 electrolyte exhibiting the highest proportion of SSIP at 37.57%, compared to 14.21% and 1.81% for TFAD/LiClO4 and TFAD/LiTFSI systems, respectively. Conversely, aggregate (AGG) formation is most pronounced in the LiTFSI/TFAD system (57.07%), followed by TFAD/LiClO4 (45.28%) and TFAD/LiPF6 (9.37%) (Figures 2F and S16). These results provide important insight into the origin of viscosity differences among distinct DES systems. Previous studies have reported that an increased fraction of ionic AGG correlates with elevated viscosity and diminished ionic diffusivity in ionic liquids and concentrated electrolytes [44, 45]. Consistent with these findings, our results confirm the same trend in DES systems. The viscosities of TFAD/LiPF6, TFAD/LiClO4, and TFAD‐LiTFSI were measured to be 13.8, 30.7, and 42.8 cP, respectively, in clear positive correlation with the fraction of AGG (Figure S17). Notably, the TFAD/LiPF6 system demonstrates markedly superior transport properties, with Li+ and PF6 − diffusion coefficients of 3.34 × 10−7 and 3.61 × 10−7 cm2s−1, respectively, approximately an order of magnitude higher than those determined for the TFAD/LiClO4 and TFAD/LiTFSI electrolytes (Figure S18). To further probe ion transport within the electrolyte, temperature‐dependent EIS measurements were carried out (Figure S19). The ionic conductivity of the DES as a function of temperature follows a clear Arrhenius relationship (Figure 2G), indicating a thermally activated transport mechanism with a single activation energy. Such linear behavior suggests that ion migration is not dominated by sluggish, multi‐step relaxation processes but is governed by relatively uniform activation barriers. While PF6 − anions dynamically exchange between SSIP, CIP, and AGG states, the Arrhenius dependence implies that these interactions are highly reversible, allowing PF6 − to maintain effective contribution to bulk conductivity. This interpretation reconciles the spectroscopic and simulation evidence with transport measurements, confirming that the favorable ion transport in LiPF6/TFAD originates from the coexistence of dissociated ions and transiently coordinated structures. Having established the favorable ion transport characteristics of the eutectic anolyte, we next turned to the selection of a suitable catholyte phase. To identify a suitable catholyte for the biphasic system, halogenated solvents were first evaluated in terms of their electrochemical stability and ability to dissolve the supporting electrolyte. Saturated solutions of TBAPF6 were prepared in DCM, 1,2‐dichloroethane (DCE), and 1,2‐dichlorobutane (DCB), and their electrochemical stability windows were measured (Figure S20). All three solvents exhibited sufficient oxidative stability to support high‐voltage operation. However, DCM demonstrated the most favorable combination of high salt solubility and low viscosity, suggesting superior ion transport properties among the three. To further assess their ionic conductivity, EIS measurements were performed on the individual saturated TBAPF6 solutions. Among them, the DCM‐based solution showed the lowest resistance of 107.5 Ω cm2, indicating the highest ionic conductivity. This trend remained consistent when each solvent was combined with the TFAD/LiPF6 (5:1) eutectic to form a biphasic system. As shown in the Nyquist plots (Figure S21), the DCM||TFAD/LiPF6 configuration exhibited the lowest overall impedance, including both potential bulk and interfacial resistance (82.3 Ω cm2), this finding confirms not only the superior ion accessibility across the liquid–liquid interface but also suggests that the DES phase provides a comparable ion transport pathway compared to the organic solvent phase. The physical phase behavior of these biphasic systems was also examined. As shown in Figure S22, all solvent‐DES pairs formed clear and stable phase separations without visible emulsification. Saturated DCM, with a density of ∼1.235 g·cm−3, consistently floated above TFAD/LiPF6 (1.327 g·cm−3), ensuring robust vertical stratification and facile separation between the catholyte and anolyte. Taken together, these results confirm that DCM is the optimal catholyte solvent, offering the best electrochemical performance and phase stability in both single‐phase and biphasic systems for membrane‐free Li–organic hybrid flow battery operation. Selection and Stability Test of Redox‐Active Materials To ensure long‐term operational stability in a biphasic RFB, the redox‐active species must not only possess high solubility and great electrochemical reversibility but also remain strictly confined to the catholyte phase, without crossover into the anolyte. This phase confinement is particularly critical in membrane‐free systems, where uncontrolled diffusion across the liquid–liquid interface can lead to capacity loss and performance degradation [46]. A series of organic molecules were screened for their suitability, including phenothiazine derivatives, anthraquinone derivatives, cyclopropenium derivatives, tetrathiafulvalene derivatives, ferrocene, and TEMPO‐based radicals [47, 48, 49, 50, 51, 52]. These candidates were selected for their established redox activity in nonaqueous media and structural tunability. The solubility of each compound was first evaluated in both the DES (TFAD/LiPF6) and DCM‐based catholytes. As summarized in Table S2 and Figure S23, most species showed poor solubility in the DES and moderate to good solubility in DCM. However, several candidates exhibited partial or full crossover into the DES phase when the two layers were brought into contact for 12 h. In particular, small and highly polar molecules such as ferrocene, TEMPO, and anthraquinone derivatives were prone to diffusing into the bottom phase, compromising the phase isolation. Among all tested species, only long‐alkyl‐chain‐substituted phenothiazines, specifically C8‐PTZ and C18‐PTZ exhibited strict confinement to the DCM layer over time, as revealed by CV scans of the DES layer (Figures S24 and S25). The extended alkyl chains likely increase hydrophobicity and reduce miscibility with the polar DES, thereby promoting preferential partitioning into the catholyte. Between the two, C8‐PTZ was chosen as the optimal redox‐active compound due to its superior solubility in DCM (miscible, with a neat‐liquid equivalent concentration of approximately 3.39 M) compared to C18‐PTZ (∼ 0.47 M), which ensures a higher theoretical capacity and improved energy density. The redox behavior of C8‐PTZ was further characterized using low‐scan‐rate CV in saturated TBAPF6/DCM. As shown in Figure S26, the compound exhibited a pair of sharp and symmetric redox peaks with peak separation of ∼ 77 mV, indicative of a reversible one‐electron redox process [53]. The peak currents remained stable over 400 consecutive cycles, and no new peaks or significant decay in signal intensity was observed (Figure S27). These results confirm that C8‐PTZ undergoes reversible redox transformation with no apparent degradation or side reactions under the experimental conditions. Taken together, the excellent phase selectivity, high solubility, and robust redox performance of C8‐PTZ make it a highly suitable catholyte for long‐term membrane‐free biphasic Li–organic hybrid flow battery operation. Dynamics The redox kinetics and diffusion characteristics of C8‐PTZ in saturated TBAPF6/DCM were systematically studied using CV and rotating disk electrode (RDE) techniques (Figure 3). In the CV measurements conducted at scan rates ranging from 10 to 750 mV s−1 (Figure 3A), C8‐PTZ exhibited a pair of well‐defined and symmetric redox peaks, and the anodic‐to‐cathodic peak current ratios (ipa/ipc) were consistently close to 1 across all scan rates (Figure S28), indicating high electrochemical reversibility. All voltammograms with peak separations below 200 mV were considered to represent reversible or quasi‐reversible electron transfer processes and were subsequently used for diffusion coefficient calculation (Randles–Ševčík equation). The peak current densities were linearly proportional to the square root of the scan rate (Figure 3B), confirming that the redox process is diffusion‐controlled under these conditions. RDE measurements were then conducted at rotation rates ranging from 100 to 2500 rpm (Figure 3C). The limiting current increased proportionally with rotation speed, further confirming that the redox process is governed by diffusion. From the slope of the current–rotation plot (Figure 3D), the diffusion coefficient of C8‐PTZ was estimated to be 8.00 × 10−5 cm2 s−1. Koutecký–Levich analysis (Figure 3E) revealed a standard rate constant of 1.05 × 10−2 cm s−1, suggesting that C8‐PTZ undergoes fast and efficient charge transfer at the electrode interface. Tafel fitting (Figure 3F) yielded a moderate slope of 70.2 mV·dec−1, further supporting rapid kinetics with minimal interfacial resistance [54]. Overall, these results confirm that C8‐PTZ possesses favorable diffusion and kinetic characteristics for use in membrane‐free nonaqueous flow batteries. The efficient electron transfer and stable redox response across different regimes ensure compatibility with both static and dynamic operations. (A) Cyclic voltammograms of 1.0 mM C8‐PTZ in DCM at various scan rates from 10 to 750 mV s−1. (B) Peak current density vs. square root of scan rate for oxidation and reduction processes. (C) Linear sweep voltammetry plots obtained at different rotation rates of the rotating disk electrode. (D) Peak oxidation and reduction current density vs. square root of angular velocity of first electron. (E) Koutecký–Levich curve (current−1 vs ω−1/2) at different reduction overpotentials of the first electron. (F) Tafel plot constructed using the current response and overpotentials. Cycling Performance Under Static Conditions The cycling performance of the membrane‐free biphasic Li–organic hybrid flow battery was evaluated under static conditions at ambient temperature and pressure, with a current density of 2.5 mA·cm−2. The cell employed a lithium metal anode, a eutectic TFAD/LiPF6 electrolyte as the bottom‐phase anolyte, and C8‐PTZ dissolved in saturated TBAPF6/DCM as the top‐phase catholyte. Two concentrations of C8‐PTZ (0.1 and 0.5 M) were tested to assess the impact of redox‐active species concentration on battery stability. In the 0.1 M C8‐PTZ system, a TFAD:LiPF6 molar ratio of 5:1 was employed. Due to the low concentration of redox‐active species, the extent of Li+ extraction from LiPF6 during cycling was limited, resulting in minimal changes to the electrolyte composition. The eutectic phase remained in a stable liquid state throughout charge/discharge cycles, ensuring adequate conductivity and %fluidity. As shown in Figure 4A, the battery using 0.1 M C8‐PTZ delivered an initial discharge capacity of ∼0.332 mAh and retained 92.4% (∼0.307 mAh) of its capacity after 100 cycles (∼16 h of operation), with an average per‐cycle capacity retention of 99.92%. The Coulombic efficiency remained stable at 97.7%, indicating highly reversible redox behavior. The charge/discharge profiles (Figure 4B) showed moderate polarization growth, which likely accounts for the observed capacity fade. The discharge voltage plateau was centered at approximately 3.6 V, consistent with the redox potentials determined from CV (Figure S29). While the CV curves of C8‐PTZ before and after cycling (Figure 4C) exhibited a slight decrease in peak current but retained nearly identical redox potentials, suggesting no major structural degradation of the redox‐active species. As illustrated in Figure S30, voltage efficiency gradually decreased over time due to increasing polarization, whereas energy efficiency remained above 80%, demonstrating overall performance stability. In addition, CV of the TFAD/LiPF6 phase after prolonged cycling (Figure S31) showed no detectable redox activity, further confirming the electrochemical stability of the anolyte and its effective suppression of active species crossover in this membrane‐free configuration. A 0.1 M Li/C8‐PTZ static battery at a current density of 2.5 mA cm−2 (A) Cycling performance (B) Galvanostatic charge/discharge profiles and (C) Cyclic voltammograms of catholyte before and after cycling; A 0.5 M Li/C8‐PTZ static battery at a current density of 2.5 mA cm−2 (D) Cycling performance (E) Galvanostatic charge/discharge profiles and (F) Cyclic voltammograms of catholyte before and after cycling. In the 0.5 M C8‐PTZ system, a TFAD:LiPF6 molar ratio of 4.5:1 was deliberately chosen to maintain the stability of the eutectic phase throughout the charge/discharge process. At higher concentrations of redox‐active species, the progressive extraction of Li+ from the anolyte during cycling significantly alters the salt‐to‐solvent ratio. If a system of TFAD:LiPF6 5:1 ratio were used, progressive Li+ extraction during cycling would drive the system beyond its eutectic liquidus boundary (approaching an effective ratio of ∼5.8:1), thereby inducing phase destabilization, viscosity escalation, and a marked decline in ionic conductivity. The 4.5:1 formulation was optimized to avoid this issue, ensuring the system remains fully liquid across the entire state‐of‐charge range. Figures 4D and S32 present the cycling performance of the 0.5 M C8‐PTZ system. The battery delivered a peak discharge capacity of approximately 1.62 mAh, corresponding to a theoretical capacity utilization of 92.9%. Over 100 cycles spanning 61 h, the average discharge capacity was approximately 1.49 mAh, corresponding to an average theoretical capacity utilization of 85.6%, with an average Coulombic efficiency of 97.6% and an average per‐day capacity retention of 99.45%. As shown in Figure 4E, distinct voltage plateaus were maintained over extended cycling, a noticeable increase in polarization was observed compared to the 0.1 M system, likely due to increased viscosity and concentration‐dependent mass transport limitations. Notably, the CV profiles of C8‐PTZ before and after cycling (Figure 4F) exhibited minimal shifts in peak potential and only moderate current attenuation, suggesting that the redoxmers retained their electrochemical activity and structural stability throughout the prolonged operation. Despite the higher viscosity of the 0.5 M catholyte, the battery achieved robust energy performance. Moreover, the CV of the TFAD/LiPF6 anolyte after extended cycling (Figure S33) showed no detectable redox peaks, confirming an excellent crossover suppression property of the bi‐phase battery. Battery Performance Under Flow Conditions To maintain phase stability throughout extended operation, the 0.5 M C8‐PTZ flow battery employed a TFAD:LiPF6 ratio of 4:1 in the anolyte, consistent with the design rationale used in static configurations. This slightly more concentrated salt composition prevents the electrolyte from exceeding the eutectic liquid‐phase boundary during continuous Li+ depletion and re‐deposition, which is particularly important under flow conditions, where extended cycling and higher redox throughput can amplify compositional drift. Under galvanostatic cycling at 2.5 mA·cm−2, the flow battery demonstrated excellent capacity retention over 100 cycles (Figure 5A). The initial discharge capacity was ∼5 mAh, and the cell retained 92.6% of its initial discharge capacity after 10 days of continuous operation. The average per‐cycle and per‐day capacity retention values were 99.91% and 99.25%, respectively. This high stability, nearly matching that of the static system, indicates that mechanical shear and convective mixing do not disrupt phase integrity in the biphasic configuration. The charge/discharge profiles (Figures 5B and S34) remained symmetric, though gradual increases in overpotential were observed, likely due to growing interfacial resistance and moderate mass transport limitations. CV analysis before and after cycling (Figure 5C) showed minimal shift in redox peak potentials and a ∼5% reduction in peak current, suggesting that the redox‐active C8‐PTZ remained largely intact. This decline may partially account for the observed capacity fade. The underlying degradation mechanism, however, is not yet fully understood. Post‐cycling analysis of the anolyte revealed no detectable C8‐PTZ crossover (Figure S35), and broad‐range CV investigations (Figure S36) did not reveal any new electroactive species. Beyond the redox‐active species, performance degradation is also closely linked to cumulative imbalances in PF6 − transport, which arise because of persistent side reactions. Although the Coulombic efficiency remained above 97%, such values still imply a measurable degree of irreversibility. Over extended cycling, the mismatch between the forward and reverse fluxes of PF6 − anions leads to progressive compositional drift in the anolyte. This imbalance, confirmed by post‐cycling NMR (Figure S37), results in reduced ionic conductivity, as evidenced by increased bulk resistance in EIS measurements (Figure S38). These findings collectively elucidate the underlying causes of the increasing polarization trend observed (Figure 5B). A 0.5 M Li/C8‐PTZ flow battery at a current density of 2.5 mA cm−2 (A) Cycling performance (B) Galvanostatic charge/discharge profiles and (C) Cyclic voltammograms of catholyte before and after cycling. Rate performance and power characteristics of the membrane‐free redox flow battery. (D) Galvanostatic charge/discharge profiles of the 0.5 M Li/C8‐PTZ biphasic flow battery at current densities of 1.0–5.0 mA cm−2. (E) Discharge capacity and efficiencies of the Li/C8‐PTZ battery at various current densities. (F) Power density at room temperature under 50, 75, 100% SOC. (G) Open‐circuit voltage (OCV) and area‐specific resistance (ASR) versus state of charge. (H) Cost breakdown of the Li/C8‐PTZ membrane‐free biphasic redox flow battery. (I) Cost of the Li/C8‐PTZ membrane‐free Li–organic hybrid flow battery compared to a conventional all‐vanadium redox flow battery. Following the durability evaluation, the rate capability of the same flow cell was assessed across current densities ranging from 1.0 to 5.0 mA·cm−2 (Figure 5D). At lower currents (1.0–2.0 mA·cm−2), the cell achieved near‐theoretical capacities with minimal voltage losses. At higher current (5.0 mA·cm−2), capacity dropped significantly and polarization became more pronounced, indicating increased resistance and slower ion transport in both phases. Nonetheless, the charge/discharge curves remained symmetrical, implying that electron‐transfer kinetics were not the limiting factor. Figure 5E shows the discharge capacity, Coulombic efficiency, voltage efficiency, and energy efficiency of the battery at different current densities. Coulombic efficiency remained close to 100% throughout, while voltage efficiency decreased from approximately 90% to 74% as current density increased, resulting in Energy efficiency values ranging from 90% to 74%. These results indicate that mass transfer becomes the limiting factor at higher current loads. As shown in Figure 5F, the maximum power density reached 54 mW cm−2 at 100% SOC, which is among the relative values reported for nonaqueous membrane‐free flow battery systems [29, 55]. To further clarify the origin of this limitation, area‐specific resistance (ASR) and open‐circuit voltage (OCV) were examined as a function of state‐of‐charge (SOC) (Figure 5G). The high‐frequency ASR, corresponding primarily to bulk ionic resistance, remained relatively constant (varying by less than 2 Ω·cm2) across all SOC levels, which showed negligible dependence on changes in the composition of the bottom‐phase anolyte. This further suggests that the DCM phase acts as the rate‐limiting medium for ionic conduction at high frequencies. In parallel, the OCV profile as a function of SOC (also shown in Figure 5G) provides insight into the thermodynamic behavior of the redox couple. At low SOC regions, slight deviations from linearity are observed, likely due to changes in reactant concentration or increased overpotential. In the mid‐SOC range (25–80%), the voltage curve displays a quasi‐linear region, suggesting a stable reaction potential. The power output declined at lower SOC due to diminished redox potential separation. To evaluate self‐discharge behavior, OCV of the fully charged battery was monitored over 100 h (Figures S39 and S40). A slow and linear decay was observed, with an average rate of 0.540 and 0.503 mV h−1 under flow and static conditions, respectively. This minimal loss demonstrates the excellent phase stability and low crossover of C8‐PTZ under flow operation. The results confirm that the biphasic structure effectively inhibits interfacial self‐discharge and preserves long‐term energy storage integrity. Despite these promising flow‐cell results, it should be noted that the present configuration does not provide the same degree of independent energy‐power scalability as a conventional all‐liquid redox flow battery. The capacity of the positive side can be increased by increasing the catholyte volume and/or C8‐PTZ concentration, whereas the power output can be scaled mainly through the electrode area, number of cells, and flow‐field design. In contrast, the capacity of the negative side is determined by the lithium inventory. Nevertheless, the high volumetric capacity of solid metal electrodes partially mitigates this limitation, because a relatively small volume of metal can balance a much larger volume of dissolved catholyte, as exemplified by Zn‐based hybrid flow battery systems. The use of lithium metal further reduces the required negative‐electrode volume because of its high specific and volumetric capacity. Therefore, this battery architecture represents a trade‐off between complete independent scalability and increased cell voltage and energy density. Economic and Safety Analysis To preliminarily assess the economic potential of the membrane‐free biphasic system, a simplified cost estimation was performed based on the architecture and materials used in this study. The cost model follows the framework of membrane‐free RFB described in previous literature [56], which divides the total cost into three main components: the cell stack (including electrodes, current collectors, and gaskets), the electrolyte (solvents, salts, and active species), and the balance‐of‐plant components (pumps, heat exchangers, tanks, and controls). By eliminating the membrane, a significant portion of the stack cost is avoided. Furthermore, the immiscibility between the eutectic anolyte and DCM‐based catholyte removes the need for auxiliary rebalance systems. In our system, key cost‐contributing materials include collector (graphite felt and Nickel foam), salt (TBAPF6, LiPF6), and C8‐PTZ. Although the organic solvents and salts are relatively expensive compared to aqueous systems, the absence of membranes and simplified system design help reduce the overall cost. Using an adapted Regenesys‐based benchmark and assuming a storage system rated at 2 MW × 6 h, the estimated total system cost for our membrane‐free Li–organic hybrid flow battery was calculated to be approximately USD $134.8 per kWh (see Table S3 and Supporting Information for detailed information), which is lower than traditional vanadium RFBs (USD $251.9 per kWh) in this model (Figure 5I) [56]. The reduction arises mainly from the exclusion of membrane materials and associated simplifications. While further scale‐up, material optimization, and recycling schemes could improve cost‐effectiveness, this result highlights the economic viability of membrane‐free designs for stationary energy storage applications. Preliminary safety evaluations were also conducted, focusing on the flammability of the electrolyte components. As shown in Figure S41, open‐flame ignition tests were performed on representative samples of the TFAD/LiPF6 (5:1) DES, saturated TBAPF6/DCM, and conventional acetonitrile‐based nonaqueous electrolytes. The DES and TBAPF6/DCM strip failed to ignite even after 2 s of direct flame exposure (Figure S41B,E), confirming its excellent flame resistance. In contrast, a standard acetonitrile electrolyte ignited instantly and sustained combustion for more than 15 s (Figure S41I), demonstrating significantly higher flammability. These results highlight the intrinsic fire‐retardant nature of the TFAD/LiPF6 eutectic, likely attributed to the low volatility of the DES and the high fluorine content in TFAD, which collectively suppress solvent vaporization and flame propagation. Similar flame‐resistant behavior has also been observed in high‐temperature and combustion tests of structurally related trifluoroacetamide derivatives, further corroborating the flame‐retardant characteristics of this eutectic system [57, 58]. The nonflammable characteristics of both electrolyte phases substantially reduce the risk of thermal runaway and fire hazards during operation, offering a significant safety advantage over traditional organic electrolytes. This intrinsic safety, when combined with membrane‐free simplification and cost reduction, strengthens the practical appeal of biphasic Li–organic hybrid flow batteries for stationary applications. These ignition tests demonstrate the excellent flame resistance of both electrolyte phases and indicate a reduced risk of electrolyte combustion. However, we acknowledge that electrolyte flame resistance does not eliminate the intrinsic safety concerns associated with lithium metal, such as dendrite formation and internal short circuits. Further optimization of the lithium metal interface and deposition behavior will therefore be required for practical applications. Conclusion In this study, we designed and demonstrated a high‐voltage membrane‐free nonaqueous Li–organic hybrid flow battery based on a biphasic electrolyte system that combines the advantages of DESs and halogenated organic media. A comprehensive screening of DES candidates identified the eutectic mixture of TFAD and LiPF6 at a 5:1 molar ratio as the most promising anolyte. This DES exhibited an electrochemical stability window extending up to 4.81 V, a low viscosity, and a high ionic conductivity of 8.59 mS·cm−1, alongside excellent phase stability and chemical compatibility with lithium metal. Its favorable physical properties, including relatively high density, allowed it to function effectively as the bottom‐phase electrolyte in a biphasic system. DCM was selected as the catholyte solvent based on its immiscibility with the DES, low viscosity, high salt solubility, and low interfacial resistance. Among several redox‐active species evaluated, long‐alkyl‐chain‐substituted phenothiazines, particularly C8‐PTZ, demonstrated outstanding solubility, redox reversibility, and strict confinement to the catholyte phase. Electrochemical analysis confirmed that C8‐PTZ exhibits fast electron‐transfer kinetics and diffusion‐controlled behavior, with a standard rate constant on the order of 10− 3 cm·s−1 and a Tafel slope of ∼70 mV·dec−1, indicating efficient redox kinetics. Static‐cell cycling at C8‐PTZ concentrations of 0.1 and 0.5 M showed high Coulombic efficiencies (>97%) and good voltage and energy efficiencies. The 0.1 M cell retained 92.4% of its initial discharge capacity after 100 cycles, while the 0.5 M cell achieved an average theoretical capacity utilization of 85.6% and an average per‐day capacity retention of 99.45% over 100 cycles. Flow cell tests further validated the robustness of the biphasic design, maintaining over 92% of the original capacity and showing negligible redox degradation or crossover, as confirmed by CV analysis. Importantly, rational control of the initial TFAD:LiPF6 ratio (4.5–5.0:1) enabled tolerance to Li+ fluxes during operation, maintaining the DES in a stable liquid state without compromising conductivity or interfacial performance. This work highlights a practical and effective strategy for constructing membrane‐free Li–organic hybrid flow battery by leveraging the unique solvation behavior and tunable phase properties of DES‐based electrolytes. By integrating electrochemical stability, ionic conductivity, phase compatibility, and redox performance into a single system, we demonstrate a path forward for next‐generation, scalable, and low‐cost energy storage solutions suitable for stationary grid applications. Future work will expand on redox couple design, flow field engineering, and long‐duration performance metrics to further bridge the gap between laboratory feasibility and practical deployment. Author Contributions Xiao Wang: writing – review and editing, writing – original draft, conceptualization, methodology. Bindu Dahal: methodology, writing – original draft, writing – review and editing. Yuhang Tang: investigation, software, methodology. Jack J. McGrath: investigation. Rabin Siwakoti: investigation. Feng Wang: supervision, methodology, funding acquisition, software. Jun Cheng: funding acquisition, methodology, software, supervision. Jianbing “Jimmy” Jiang: conceptualization, methodology, writing – review and editing, funding acquisition. Conflicts of Interest The authors declare no conflict of interest. Supporting information Supporting File: anie73742‐sup‐0001‐SuppMat.pdf. Acknowledgments This study was supported by the US National Science Foundation under Grant Number. CBET‐2112798. Contributor Information Feng Wang, Email: fengwang@xmu.edu.cn. Jun Cheng, Email: chengjun@xmu.edu.cn. Jianbing “Jimmy” Jiang, Email: Jianbing.jiang@uc.edu. Data Availability Statement The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions. References Associated Data Supplementary Materials Supporting File: anie73742‐sup‐0001‐SuppMat.pdf. Data Availability Statement The data that support the findings of this study are available on request from the corresponding author. The data are not publicly available due to privacy or ethical restrictions.