Fossil fuels (coal, oil, and natural gas), which account for over 80% of primary energy consumption, still dominate the global energy mix. Their combustion not only depletes finite fossil resources but also contributes to climate change and air pollution^[1,2]. Renewables like wind and solar have grown rapidly, yet their output is inherently variable: the sun sets, the wind lulls, and grid operators are left scrambling to balance supply and demand, often wasting excess power in the process^[3,4]. Redox flow batteries (RFBs) offer a particularly promising solution to the above challenges^[5,6]. Unlike conventional closed-cell batteries which store energy directly in the electrode materials, RFBs store energy in liquid electrolytes. Moreover, these electrolytes are contained in separate external tanks. This architecture means that power and energy capacity become independent of each other. This feature is especially valuable for grid-scale energy storage. RFBs also bring other practical benefits. For example, they are intrinsically safe, tolerate deep discharging, and typically have a long cycle life^[7-10]. Among the various RFB chemistries, the all-vanadium redox flow battery (VRFB) has been the most commercially advanced with several megawatt-scale installations already in operation worldwide^[11-14]. Despite this progress, the VRFBs still face notable limitations that hinder wider deployment. Vanadium is a relatively scarce and costly metal, which causes higher system costs. Meanwhile, the energy density which is determined by the concentration of active species is another issue as the current vanadium electrolytes only reach about 25 Wh L^-1. Furthermore, the strongly acidic electrolyte creates a harsh environment that can promote the degradation of electrodes, membranes, bipolar plates, and other cell components, thereby compromising the long-term durability of VRFBs^[15,16]. These drawbacks have motivated considerable research into alternative chemistries that are affordable, more energy-dense, and less corrosive to cell hardware^[17,18].

This is where the alkaline ferrocyanide/polysulfide redox flow battery (Fe/S RFB) enters the picture. Its active materials, ferrocyanide and polysulfide, are cheap, abundantly available, and dissolve readily in alkaline electrolyte (up to 1.0 M or higher)^[19-22]. The trouble is, the S₂^2-/S^2- couple at the anode suffers from a large kinetic barrier, causing sluggish charge transfer and severe polarization that cripple efficiency and cycling stability^[23,24]. Therefore, the development of efficient electrocatalytic materials to accelerate the charge-transfer process of the sulfide redox pair is of paramount importance for advancing this battery system^[25].

Recent efforts have explored a variety of transition-metal sulfides (e.g., CoS, WS₂, and MoS₂) and their carbon-based composites as electrocatalysts for modifying carbon felt (CF) electrodes to boost polysulfide conversion^[26-28]. For instance, Liu and co-workers modified CF with 1T-MoS₂, achieving an EE of 80.82% at a current density of 20 mA cm^-2 over 2,500 cycles^[28]. Despite these encouraging results, many of the reported catalytic materials suffer from drawbacks such as complicated preparation procedures, high raw-material costs, or limited density of active sites, which collectively impede their simultaneous fulfillment of high activity, long-term durability, and economic viability for practical applications.

To address these challenges, we herein propose a novel and sustainable strategy that repurposes waste asphalt, which is an abundant industrial byproduct, as a carbon precursor, combined with electro-deposited copper sulfide (CuS), to construct a highly active interface on CF. In our approach, waste asphalt mixed with KOH was uniformly coated onto CF and subsequently subjected to high-temperature pyrolysis. The KOH in our approach serves a dual role. It creates abundant defect sites through chemical activation and markedly improves the hydrophilic character of the carbon surface, thus facilitating electrolyte penetration. Following pyrolysis, copper was electrodeposited onto the pyrolytic carbon layer. The resulting material was then immersed in K₂S solution to convert the deposited copper into CuS nanoparticles. This yields the waste-asphalt pyrolytic carbon and electrodeposited CuS co-modified CF (denoted as CuS-WA/CF).

Characterization results confirm that CuS nanoparticles are uniformly anchored on the carbon fibers. The specific surface area is significantly enlarged. Wettability by the alkaline electrolyte is also greatly enhanced. Electrochemical tests show that the CuS-WA/CF electrode displays superior activity toward the S₂^2-/S^2- reaction. It reduces the electrochemical polarization and promotes interfacial charge transfer.

We set the anode to operate on the reversible S₂^2-/S^2- redox couple. With the CuS-WA/CF electrode, the alkaline Fe/S RFB shows much less polarization and reaches a peak power density of 259.2 mW cm^-2. Under long cycling at 40 mA cm^-2, it delivers an average energy efficiency (EE) of 77.7% and an average voltage efficiency (VE) of 78.0%. More remarkably, under the capacity-limited short-cycle protocol, the cell completed 25,000 cycles with an average EE of 84.0%. These results confirm the practical potential of the CuS-WA/CF electrode for advanced alkaline Fe/S RFBs.

We electrodeposited copper onto CF that had been pre-impregnated with waste-asphalt pyrolytic carbon. A subsequent sulfidation step converted the copper into CuS to yield the CuS-WA/CF electrode [Figure 1A]. SEM images reveal the surface morphology changes. Pristine CF shows smooth, bare fibers with a limited number of active sites for redox reactions [Figure 1B and C]. Direct copper plating on bare CF (yielding Cu-CF) produces large, smooth copper crystals at sparse nucleation sites [Supplementary Figure 1]. This morphology reduces the specific surface area and limits accessible active sites for polysulfide conversion. The waste-asphalt-derived carbon interlayer changes this picture entirely. It promotes dense and uniform Cu nucleation during electrodeposition. After K₂S sulfidation, the deposited Cu particles are converted to CuS without substantial morphological disruption. As can be seen in Figure 1D and E, the resulting CuS-WA/CF electrode exhibits a uniform, fine-grained CuS layer on the carbon fibers. The elemental maps [Figure 1F and G] show the distributions of Cu and S, respectively, within CuS coating.

XRD patterns were used to examine the crystal structure of the modified electrode [Figure 2A]. The XRD pattern of CuS-WA/CF exhibits well-defined diffraction peaks at 29.2°, 31.7°, 32.7°, and 47.8°, which can be readily indexed to the (102), (103), (006), and (110) planes of hexagonal covellite CuS (Powder Diffraction File No. 01-073-6495). The absence of impurity peaks suggests that the sulfidation reaction proceeded to completion and that phase-pure CuS was successfully deposited onto the CF substrate. We further analyzed the surface chemical composition of CuS-WA/CF using XPS spectroscopy. The survey spectrum in Figure 2B shows clear signals for carbon, sulfur, and copper. The carbon signal originates from both the underlying CF support and the waste-asphalt-derived carbon interlayer, whereas the S and Cu signals arise from the outermost CuS coating. The presence of all three elements is consistent with the expected structure of the electrode. Raman spectroscopy was employed to assess changes in carbon defect density upon modification [Figure 2C]. Both samples display the characteristic D band (~1,350 cm^-1) and G band (~1,580 cm^-1) of carbon materials. For pristine CF, the intensity ratio I_D/I_G is calculated to be 1.12, whereas for CuS-WA/CF this ratio increases to 1.42. The higher I_D/I_G value points to a lower degree of graphitization and a greater abundance of structural defects in the modified electrode. Such defects are generally beneficial for electrochemical performance, as they can provide additional active sites for reaction and may also contribute to an expanded electrochemically active surface area. To quantify the change in surface area, we performed nitrogen adsorption-desorption measurements [Figure 2D]. According to BET results, CuS-WA/CF reaches a specific surface area of 80.46 m² g^-1, which is almost nine times higher than the value of pristine CF (9.55 m² g^-1). This obvious increase illustrates that the modification creates a more porous electrode structure, which should improve electrolyte contact and ion transport during cell operation.

As reported in an earlier study, wettability also plays a key role in electrode performance^[37]. A hydrophilic surface helps the electrolyte penetrate fully and contact well with the active sites. Such good contact boosts fast electrochemical kinetics. Herein, the contact angle measurements provide direct evidence. Pristine CF shows a contact angle of 114° [Supplementary Figure 2A], indicating the hydrophobic feature. The CuS-WA/CF electrode is almost completely wetted by water, with a contact angle approaching 0° [Supplementary Figure 2B], confirming its hydrophilic character. This dramatic improvement in wettability is attributed to the combined effects. One effect is from KOH activation during high-temperature pyrolysis, which introduces oxygen-containing functional groups. The other effect is the roughness of the carbon surface. These changes also promote electron transfer and lower the interfacial resistance.

CV was employed to evaluate the electrochemical performance of the pristine CF and CuS-WA/CF electrodes toward polysulfide redox reactions. It is known that multiple polysulfide species are present in alkaline electrolytes. But well-defined redox peaks mainly corresponding to the S₂^2-/S^2- couple are observed in the CV profiles. As presented in Figure 3A, the peak separation (ΔE_p) for the pristine CF was as large as 787.9 mV. CuS-WA/CF reduces this separation to only 77.5 mV. It is worth mentioning that the CuS-WA/CF electrode also shows additional redox peaks in the same potential range. This demonstrates its electrocatalytic activity toward not only the S₂^2-/S^2- couple but also other polysulfide conversions. This enhanced electroactivity was consistently observed across various scan rates [Figure 3B and C]. Meanwhile, for both electrodes, the peak currents of the oxidation and reduction waves increased with increasing scan rate and exhibited a linear dependence on the square root of the scan rate, suggesting that the redox reactions of the S₂^2-/S^2- couple are diffusion-controlled processes [Figure 3D].

The superior conductivity and charge-transfer kinetics of CuS-WA/CF were further corroborated by EIS [Figure 3E and F, Supplementary Table 1]. The ohmic resistance (R_o) values are 1.18 and 1.15 Ω for pristine CF and CuS-WA/CF, respectively. The nearly unchanged R_o indicates comparable cell and electrolyte resistance. The charge-transfer resistance (R_ct) of CuS-WA/CF was determined to be 0.16 Ω, which is 96.9% lower than that of the pristine CF (5.11 Ω). This marked reduction in R_ct unequivocally demonstrates that the uniformly dispersed CuS particles anchored on the pyrolysis carbon skeleton promote faster interfacial charge transfer on the electrode, accelerating the oxidation-reduction kinetics of the S₂^2-/S^2- couple.

We assembled two sets of alkaline Fe/S RFBs to test the CuS-WA/CF electrode. One used pristine CF as the anode. The other used CuS-WA/CF. All other components and conditions were kept identical. The charge-discharge profiles were recorded at 40 mA cm^-2 [Figure 4A]. The CuS-WA/CF cell charges at a notably lower voltage than the pristine-CF cell. The discharge voltages of the two cells are nearly the same. This difference in charging overpotential means that CuS-WA/CF substantially reduces polarization. The polarization and power density curves confirm this [Figure 4B]. The CuS-WA/CF cell reaches 259.2 mW cm^-2 in peak power density. The pristine-CF cell reaches only 160.6 mW cm^-2. We used GITT to examine the overpotential during cycling [Figure 4C]. CuS-WA/CF significantly cuts the voltage loss. At 40 mA cm^-2, the CuS-WA/CF cell deviated from the equilibrium potential by only 83 mV, which is considerably lower than the 354 mV observed for the pristine-CF-based cell.

Rate capability tests covered 20 to 100 mA cm^-2. As shown in Figure 4D, the pristine CF and CuS-WA/CF cells exhibit comparable Coulombic efficiencies (CE) over the current density range of 20-100 mA cm^-2, with only minor differences between them. In contrast, the CuS-WA/CF cell delivers markedly higher energy efficiencies, indicating that its performance advantage primarily arises from reduced polarization. At 20 mA cm^-2, the CuS-WA/CF cell gives an average EE of 89.0%, substantially higher than the 68.9% obtained with the pristine-CF anode. At 40 mA cm^-2, CuS-WA/CF maintains 85.3%, while pristine-CF drops to 58.3%. At 60 mA cm^-2, the superiority of the CuS-WA/CF electrode becomes more pronounced, with an average EE of 79.9% compared to only 49.4% for pristine-CF. When the current density was increased to 80 and 100 mA cm^-2, the difference between the two electrodes became even more pronounced. The CuS-WA/CF cell still holds up reasonably well, with EEs of 74.3% and 68.9%, while the pristine CF cell drops sharply to 42.8% and 37.9%. The charge-discharge profiles demonstrate the underlying reason for this divergence [Figure 4E and F]. The CuS-WA/CF cell maintains a low charging overpotential and a stable discharge plateau. In contrast, the pristine CF cell experiences severe voltage losses under high-current operation. Therefore, the CuS-WA/CF electrode effectively avoids such voltage losses. It is worth noting that when the current density returns to 20 mA cm^-2, the CuS-WA/CF cell recovers to 88.8% EE, nearly matching its initial performance. But the pristine-CF cell only recovers to 65.1%. This distinct recovery behavior is a good indication that the CuS-WA/CF electrode is more electrochemically stable and reversible over a broad operating window.

We began our evaluation with a relatively dilute ferrocyanide catholyte (0.1 M [Fe(CN)₆]^4-), which served as a baseline condition for comparing the two electrode materials. At 40 mA cm^-2, the cell with the CuS-WA/CF anode reached a maximum EE of 84.6% and a VE of 85.0%. The improvements are 19.0% and 19.5%, respectively, over the pristine CF cell [Figure 5A and B]. As cycling continued, we observed a gradual decline in performance. Closer examination suggested that this decay was primarily associated with the crossover of polysulfide species from the anolyte through the membrane into the catholyte compartment, where they eventually precipitated as insulating sulfur, although electrode polarization may also have contributed to the performance loss. This observation motivated us to design a cycling protocol to distinguish electrolyte degradation from possible electrode degradation. Taking advantage of the flow-battery architecture, we replaced the electrolyte solution without disassembling the cell. In this way, the electrode remained intact and could be tested again with fresh electrolytes. When the EE of the CuS-WA/CF cell dropped to 69.0% during operation, the electrolyte-renewal procedure, including redissolution of the precipitated sulfur (details in the Section “Recovery of cell performance via electrolyte renewal”), was performed. As a result, the EE jumped back to 81.2%. Over the entire 10,000 cycles, the CuS-WA/CF cell maintained average values of 77.7% for EE and 78.0% for VE. The substantial recovery in cell performance after electrolyte replacement suggests that electrolyte degradation was a major contributor to the performance decay and that the CuS-WA/CF electrode maintained good stability under the conditions examined, although electrode-related degradation cannot be completely excluded. We compared the present system with recently reported polysulfide-based flow batteries. As summarized in Supplementary Figure 3, the CuS-WA/CF cell exhibits outstanding power output with notably prolonged cycling durability^{[22,24,28,38-41]}.

Then we raised the catholyte concentration to 0.8 M [Fe(CN)₆]^4-. It is known that higher concentration gives better energy density. However, it also makes mass transport and reaction kinetics more challenging. The electrode must operate under more demanding conditions. The CuS-WA/CF electrode performed well at this elevated concentration. The cell with CuS-WA/CF as the anode delivered an average EE of 79.7% and a VE of 80.1% over 1,100 consecutive cycles [Figure 5C and D]. The pristine CF cell, on the other hand, performed poorly under the same conditions. Its average EE and VE were only 49.9% and 50.1%, respectively, and the cell failed completely after just 500 cycles.

Encouraged by these results, we pushed the CuS-WA/CF electrode harder using a capacity-limited short-cycle protocol. This test was designed to accelerate aging and reveal the electrode’s durability. The cell was first charged to a 50% SOC. Each cycle then delivered and accepted a fixed capacity of 10 mAh [Figure 5E]. As cycling continued, both the charge and discharge voltages gradually drifted downward. We attribute this decay to self-discharge caused by crossover of active species through the membrane. When the discharge voltage fell below 0.4 V, we recharged the cell to 50% SOC and resumed testing. Under this capacity-limited protocol, the CuS-WA/CF electrode lasted for 25,000 cycles with an average EE of 84.0%. Considering the repeated capacity-limited charge-discharge cycles and periodic SOC-recovery steps, the observed durability is encouraging. These results demonstrate the potential of the CuS-WA/CF electrode for practical applications in alkaline Fe/S RFBs.

To investigate the origin of the enhanced activity of CuS-WA/CF toward the S₂^2-/S^2- redox couple, spin-polarized DFT calculations were performed to compare the adsorption behavior of the sulfur species on CuS-WA/CF and pristine CF. The optimized adsorption configurations on CuS-WA/CF are shown in Figure 6A. The adsorption energies of S₂^2- and S^2- on CuS-WA/CF are -1.75 and -2.22 eV, respectively, more negative than the corresponding values on pristine CF (-0.30 and -1.08 eV) [Figure 6B]. These results indicate that CuS-WA/CF interacts more strongly with both sulfur species and stabilizes them more effectively at the electrode surface. The differential charge-density maps provide further insight into these interactions. Pronounced charge accumulation and depletion are observed at the CuS-WA/CF interface, particularly around the contacts between the adsorbed sulfur species and surface Cu sites [Figure 6C]. In contrast, charge redistribution on pristine CF is considerably weaker and more localized [Figure 6D]. The stronger interfacial electronic coupling on CuS-WA/CF is favorable for electron transfer during the S₂^2-/S^2- conversion. These calculations support the experimentally observed improvement in redox kinetics and reduced polarization of the CuS-WA/CF electrode.

In summary, we have developed a CuS-WA/CF composite electrode using a three-step fabrication route: waste-asphalt-derived pyrolytic carbon coating, copper electrodeposition, and final sulfidation to CuS. The pyrolytic carbon interlayer plays a critical role in this approach. It provides abundant defect sites that promote uniform copper nucleation across the fiber surface. Without this layer, copper forms sparse, coarse crystals on bare CF. After sulfidation, the resulting CuS particles serve as active centers for S₂^2-/S^2- redox conversion.

Electrochemical tests confirm the performance improvements of CuS-WA/CF compared to pristine CF. The peak-to-peak separation in CV curves narrows from 787.9 to 77.5 mV, while the charge-transfer resistance extracted from EIS decreases from 5.11 to 0.16 Ω. In full alkaline Fe/S RFBs, the CuS-WA/CF anode delivers 259.2 mW cm^-2 in peak power density, a marked improvement over the 160.6 mW cm^-2 for the cell with pristine CF as the anode. The CuS-WA/CF electrode also enables excellent cycling stability in the Fe/S RFB. The cell maintains an average EE of 77.7% over 10,000 cycles, and 84.0% EE across 25,000 cycles at 40 mA cm^-2 under the short-cycle protocol.

Besides, this work shows that waste asphalt can be reused as a functional carbon precursor. The pyrolytic carbon interlayer strategy for tuning metal position uniformity may extend to other electrocatalytic systems. Overall, the CuS-WA/CF electrode offers a promising step toward high-performance and low-cost alkaline Fe/S RFBs for grid-scale energy storage.