A Freestanding Multifunctional Interlayer Based on Fe/Zn Single Atoms Implanted on a Carbon Nanofiber Membrane for High-Performance Li-S Batteries

Abstract

By virtue of the high theoretical energy density and low cost, Lithium–sulfur (Li-S) batteries have drawn widespread attention. However, their electrochemical performances are seriously plagued by the shuttling of intermediate polysulfides and the slow reaction kinetics during practical implementation. Herein, we designed a freestanding flexible membrane composed of nitrogen-doped porous carbon nanofibers anchoring iron and zinc single atoms (FeZn-PCNF), to serve as the polysulfide barrier and the reaction promotor. The flexible porous networks formed by the interwoven carbon nanofibers not only offer fast channels for the transport of electrons/ions, but also guarantee the structural stability of the all-in-one multifunctional interlayer during cycling. Highly dispersed Fe and Zn atoms in the carbon scaffold synergistically immobilize sulfur species and expedite their reversible conversion. Therefore, employing FeZn-PCNF as the freestanding interlayer between the cathode and separator, the Li-S battery delivers a superior initial reversible discharge capacity of 1140 mA h g⁻¹ at a current density of 0.5 C and retains a high capacity of 618 mA h g⁻¹ after 600 cycles at a high current density of 1 C.

1. Introduction

Lithium–sulfur (Li-S) batteries are extremely competitive in next-generation secondary batteries, because of high theoretical capacity of both the sulfur cathode (1675 mA h g⁻¹) and the lithium anode (3860 mA h g⁻¹) [1,2]. Also, sulfur has significant superiority in terms of natural abundance, material cost, and environmental benefits [3]. Nevertheless, the commercialization of Li-S batteries has been beset with various difficulties, including the electronic and ionic insulation of S₈ and its solid discharge products (Li₂S₂/Li₂S), the huge volume expansion after discharge, and the shuttle effect of intermediate lithium polysulfides (Li₂S_n, 4 ≤ n ≤ 8), which will trigger the active mass loss of both cathode and anode, low Coulombic efficiency, and fast capacity decay [4,5]. There are two main reasons for the shuttle effect: (i) lithium polysulfides can be soluble in an ether-based electrolyte and shuttle between the cathode and the anode [6,7]; (ii) the sluggish redox reaction kinetics between sulfur species will give rise to the accumulation of soluble polysulfides on the cathode side during cycling, thus intensifying the polysulfide shuttling driven by the concentration difference [8,9].

Extensive efforts have been invested in restraining the shuttle effect, such as sulfur host design [10,11], functional separator/ interlayer design [12,13], electrolyte additive utilization [14,15], and sparingly solvating electrolyte utilization [16,17,18], etc. Among them, the flexible freestanding interlayer can be directly placed between the cathode and separator to act as the physical barrier and the catalyst layer for simultaneously inhibiting the diffusion of polysulfide and accelerating their conversion reactions. Carbon nanofiber membranes have been used as the freestanding interlayer for Li-S batteries, by virtue of their facile preparation, good mechanical strength, and good flexibility [19,20,21,22]. Moreover, the highly conductive carbon network ensures that lithium polysulfides diffused into the interlayer maintain good electronic contact and can be further transformed to solid Li₂S₂/Li₂S during discharge, and then Li₂S₂/Li₂S deposited within the interlayer can be then reversibly converted to soluble polysulfides and return to the cathode side during the charge process. However, relying only on the weak physical confinement effect, lithium polysulfides inevitably escape from the inert carbon nanofiber interlayer during long cycles, and the intrinsic reaction kinetics of sulfur species still remains slow.

The incorporation of metal-based catalysts into the carbon substrate is among the most effective strategies to solve the above problems. These metal-based catalysts, such as metal elements [23,24], metal oxides [25,26], metal sulfides [27,28], metal nitrides [29,30,31], and heterojunction [32,33], can chemically anchor lithium polysulfides and reduce the redox reaction energy barrier, thus accomplishing the synchronous adsorption–conversion of polysulfides. It is absolutely accepted that the activity of adsorption–catalysis depends highly on the particle size of catalysts because only the surface atoms play these roles. In this regard, single-atom catalysts (SACs) show 100% atomic utilization and afford ample active sites for high-efficiency adsorption and catalysis [6,34,35]. However, the charge and discharge process of Li-S batteries involves multiple intermediates and multistep conversion reactions, where single-component SACs usually play a significant role in a particular aspect only. Zn-SACs show a prominent affinity for polysulfides because the electronegativity of Zn is obviously lower than that of various transition metals, including Fe, Co, and Ni [36]. Such a strong chemical anchoring effect is very favorable for the fast reduction in polysulfides and the uniform deposition of Li₂S [37]. Fe-SACs, as an inexpensive transition metal SAC, have been found to be effective in decreasing the decomposition barrier of insulative Li₂S and thus boosting its oxidation during the charge process [38,39]. Hence, the design of dual-metal Fe/Zn-SACs are expected to combine the advantages of both and enhance the overall electrochemical properties of Li-S batteries.

Here, the nitrogen-doped porous carbon nanofiber membrane implanted with iron and zinc single atoms (FeZn-PCNF) was fabricated by electrospinning combined with a high-temperature pyrolysis approach and directly employed as the freestanding interlayer of Li-S batteries. The porous carbon fibrous framework functions both as the interconnected network for high-efficiency electron/ion convey and as the robust skeleton to intensify the mechanical stability of the interlayer during cycling. The Fe and Zn single-atom active centers can synergistically anchor polysulfides and expedite their bidirectional conversion. Consequently, with the assistance of the multifunctional FeZn-PCNF interlayer, the Li-S batteries afford an impressive initial reversible discharge capacity of 1140 mA h g⁻¹ @ 0.5 C and retains a high capacity of 618 mA h g⁻¹ after 600 cycles at a high current density of 1C.

2. Experimental

2.1. Synthesis of Binary Zeolitic Imidazolate Frameworks (FeZn-ZIF)

The homogenous solution A and B were prepared by dissolving 1.97 g of 2-methylimidazole in 150 mL of methanol, and dissolving 1.695 g of Zn(NO₃)₂·6H₂O and 60 mg of Fe(NO₃)₃·9H₂O in 150 mL of methanol, respectively. Then, the solution A and B were rapidly mixed and continued to be stirred for 2 h. The FeZn-ZIF product was obtained after standing for 12 h at 60 °C, followed by centrifugation, washing with methanol several times, and drying in vacuum at 60 °C for 12 h.

2.2. Synthesis of FeZn-PCNF Membrane

An amount of 600 mg of FeZn-ZIF powder, 480 mg of polyacrylonitrile (PAN), and 120 mg of polyvinylpyrrolidone (PVP) were added into 7 mL of N,N-dimethylformamide (DMF). After stirring vigorously for 24 h, the dispersion was transferred into a syringe for the subsequent electrospinning. During the electrospinning process, the voltage was set to 18 kV, the feeding rate was set to 0.02 mL min⁻¹, and the distance between the tip of the spinning needle and the collector was 16–18 cm. By heating the as-made fibrous film at 280 °C for 1 h in an air atmosphere and then at 950 °C for 2 h in an N₂ atmosphere, the FeZn-PCNF membrane was obtained. For comparison, the CNF membrane was prepared without the addition of PVP and FeZn-ZIF, and the PCNF membrane was fabricated without the addition of FeZn-ZIF.

2.3. Material Characterizations

The micromorphology was investigated by a scanning electron microscope (SEM, S-4800) and a transmission electron microscope (TEM, JEM-2100F, JEOL Ltd., Tokyo, Japan). The distribution of Fe and Zn single atoms was recognized via high-angle annular dark-field scanning transmission electron microscopy (HAADF-STEM, FEI Tian Cubed Themis G2 300, Shanghai, China). The X-ray diffraction (XRD) was performed on an X’Pert PRO MPD system with Cu Kα radiation (λ = 1.5418 Å) with a scan speed of 10 min⁻¹. Pore structure was analyzed on a Micromeritics ASAP 2020 analyzer (Shanghai, China). The X-ray photoelectron spectroscopy (XPS) was detected by Thermo Scientific Escalab 250 XI (Thermo Fisher Scientific, Waltham, MA, USA) with using Al K alpha as the X-ray source of and the charge correction (C 1s at 284.8 eV) as the peak calibration procedure. Raman spectra were recorded on Renishaw RM2000 with a 512 nm laser. The mass loading of Fe and Zn atoms was measured via inductively coupled plasma atomic emission spectroscopy (ICP-AES, PerkinElmer, Shanghai, China).

2.4. Adsorption Experiments

S and Li₂S were dissolved in the electrolyte (1.0 M LiTFSI, DOL:DME = 1:1 vol%, 2.0% LiNO₃ additive, purchased from DoDo Chem, Suzhou, China) with a molar ratio of 5:1 under intensively stirring at 60 °C to obtain 1 mM of the Li₂S₆ solution. FeZn-PCNF and PCNF, with identical weights, were separately soaked into the Li₂S₆ solution. After resting for 6 h, ultraviolet-visible (UV-vis) spectra of the solution were examined on a UV-vis spectrometer (Shimadzu UV-2700, Kyoto, Japan).

2.5. Symmetrical Cell Measurements

The FeZn-PCNF and PCNF membranes were separately cut into electrode slices with 12 mm of diameter size. The symmetrical cells were then assembled using the identical electrodes as the cathode and anode and 20 μL of the Li₂S₆ solution (0.2 M) as the electrolyte. The cyclic voltammogram (CV) tests were performed on an Ivium-n-Stat electrochemical workstation at 1 mV s⁻¹ within −0.8–0.8 V.

2.6. Li-S Cell Measurements

The components of CR2032 coin cells were as follows. The cathode was fabricated by mixing the carbon nanotubes/sulfur mixture (CNTs/S), super P, and polyvinylidene fluoride with a mass ratio of 7:2:1 in the N-Methylpyrrolidone solvent. The mixed slurry was cast onto the Al foil and cut into disks with 12 mm diameter after drying. The sulfur mass loading in the CNTs/S cathode was around 1.2 mg cm⁻². The anode was lithium foil, the separator was polypropylene (PP) membrane, the electrolyte was 1.0 M LiTFSI dispersed in the DME/DOL solution (1:1 vol%) with 2.0 wt.% of the LiNO₃ additive, and the interlayer is FeZn-PCNF or PCNF membranes. The electrolyte/sulfur (E/S) ratio was 19 μL mg⁻¹. The CV curves in the voltage range of 1.7–2.8 V and the electrochemical impedance spectroscopy (EIS) in the frequency range of 0.00001–100 kHz of the as-assembled Li-S batteries were measured on the Iviumn-Stat electrochemical workstation. The galvanostatic charge and discharge (GCD) tests were conducted on a Land CT2001A system in the voltage range of 1.7–2.8 V.

3. Results and Discussion

The fabrication procedure of the FeZn-PCNF membrane is illustrated as Figure 1. Briefly, the FeZn-ZIF precursor was first synthesized via the coordination of binary metal ions (Fe²⁺ and Zn²⁺) and 2-methylimidazole ligands. The SEM images (Figure S1) indicate that FeZn-ZIF presented a typical polyhedral morphology with a uniform size. The fibrous membrane composited of PAN, PVP, and FeZn-ZIF was then fabricated by using electrospinning technology. As shown in the SEM image (Figure S2a), a multitude of nanofibers are interwoven together to form a network structure. The SEM image at higher magnification (Figure S2b) further shows a rough and uneven surface of nanofibers, indicating the successful incorporation of FeZn-ZIF polyhedrons. During the subsequent high-temperature pyrolysis process, the FeZn-ZIF polyhedrons provided a stable confined space to inhibit the agglomeration of metallic atoms, and the evaporation of partial Zn out of the FeZn-ZIF skeleton not only created many micropores in the nanofibers, but more importantly, it left behind more nitrogen sites to immobilize residual Zn and Fe atoms via the M-N_x coordination structure [36,37,40,41]. In addition, PVP molecules were degraded and thus further enriched the pore structure of carbon nanofibers. The intact fibrous membrane structure was well inherited after pyrolysis. The thickness and mass of the FeZn-PCNF interlayer are 60 μm and 1 mg, respectively. Figure S3 shows it can be folded at various angles and ultimately returned to the original state, demonstrating its good structural flexibility and mechanical stability, which are beneficial to keep the integrity during long cycling when used as the multifunctional interlayer.

The micromorphology and structure of FeZn-PCNF membrane can be observed from Figure 2. The SEM image (Figure 2a) depicts a network structure interwoven by many elongated nanofibers with rough surface, endowing FeZn-PCNF with a good flexibility and mechanical stability. Also, a large quantity of holes between the staggered nanofibers are beneficial to the immersion and transport of the electrolyte. The TEM image (Figure 2b,c) further shows the existence of many hollow polyhedral boxes, which are derived from the pyrolysis of the FeZn-ZIF precursors within nanofibers, which can function as the reservoirs for soluble polysulfides. In addition, no metal aggregation was observed in FeZn-PCNF, indicating the high dispersion of metallic atoms. The HAADF-STEM image at an atomic size (Figure 2d) further reveals that a wealth of isolated Fe/Zn atoms (seen as bright spots) are randomly distributed throughout the nanofibrous membrane, which is also confirmed by the STEM image and the corresponding EDS elemental mapping (Figure 2e–i).

A similar phase structure is demonstrated by the XRD patterns of FeZn-PCNF and PCNF. As shown in Figure 3a, there is only a broad peak at 25.4°, which is corresponding to the (002) crystal plane of non-graphitized carbon. No peaks associated with iron and zinc occur, further suggesting the amorphous state of metallic atoms. The Raman spectra (Figure 3b) of both FeZn-PCNF and PCNF exhibit two main peaks at 1354 cm⁻¹ and 1587 cm⁻¹, corresponding to the typical D band (disorder and defects in carbon) and G band (graphitic SP² carbon), respectively. Moreover, the defect degree can be revealed by the peak intensity ratio of D and G bands (I_D/I_G). The I_D/I_G value for FeZn-PCNF is 1.08, far higher than that for PCNF (0.97), which can be ascribed to a more defective structure derived from the doping of N, Fe, and Zn atoms. As can be seen from the N₂ adsorption and desorption isotherms (Figure 3c), PCNF shows type-I isotherms, which are characteristic of a micropore structure [42], but FeZn-PCNF shows type-VI isotherms, indicating both micropores and mesopores in FeZn-PCNF [43]. In contrast, CNF exhibits negligible adsorption for N₂, revealing a nonporous structure (Figure S4). The specific surface area of FeZn-PCNF, PCNF, and CNF are 237, 169, and 30 m² g⁻¹, respectively, fully demonstrating the pore-forming effects derived from the PVP pyrolysis and the incomplete evaporation of Zn from the ZIF-derived carbon skeleton. The pore size distribution curves in the inset of Figure 3c also reveal a more developed porous structure of FeZn-PCNF, which will be beneficial to exposing more active sites and accelerating ion diffusion. The surface chemical compositions of FeZn-PCNF and PCNF were identified by XPS spectra. As seen from Figure 3d, the full XPS survey spectra show that both FeZn-PCNF and PCNF contain C, O, and N elements. In addition, Zn 2p_1/2 and Zn 2p_3/2 peaks were detected in FeZn-PCNF, indicating that there is still residual Zn after pyrolysis. No elemental Fe signal was observed resulting from the limited detection depth of XPS [39]. The mass loading of Fe and Zn single atoms in FeZn-PCNF was tested as 0.78 wt.% and 1.25 wt.% via ICP-AES. The C 1s spectra of FeZn-PCNF and PCNF (Figure S5a,b) can be deconvoluted into three sub-peaks assigned to C–C/C=C, C-N, and C=O bonds [44,45]. With regard to the N 1s spectra (Figure 3e,f), the N element exists in four forms, including pyridinic N (398.0 eV), pyrrolic N (399.6 eV), graphitic N (400.7 eV), and oxidized N (402.9 eV) [46]. The proportion of various N species is displayed in Table S1. It is found that pyridinic N is significantly increased and dominates in FeZn-PCNF, which may originate from the nitrogen species in the FeZn-ZIF precursor. The abundant pyridinic N provides stable coordination sites for the monatomic dispersion of Fe and Zn, and its inherent electron-rich characteristic results in a good affinity for polar polysulfides [47,48,49]. All of the characterization results demonstrate that dual-metal Fe and Zn single atoms are evenly implanted in the porous and flexible carbon nanofiber network, which are expected to enhance the chemical adsorption for polysulfides and expedite the redox reaction during cycles.

The adsorption capability of FeZn-PCNF and PCNF membranes for soluble polysulfides was evaluated by visualized adsorption tests. As displayed in Figure 4a, the Li₂S₆ solution still shows as brownish yellow after adding PCNF, whereas the Li₂S₆ solution containing FeZn-PCNF become almost colorless. The clear supernatant of the Li₂S₆ solution after adsorption was characterized by the UV-vis spectra. Compared with the Li₂S₆ solution after being adsorbed via PCNF, the absorption peak of S₆²⁻ was hardly detected in the Li₂S₆ solution after being adsorbed via FeZn-PCNF. These comparative results manifest that the FeZn-PCNF interlayer has a potent adsorption for polysulfides, which is attributed to the comprehensive advantages including the physical adsorption by well-developed pores and the chemical anchoring by M-N_x active centers. The electrocatalytic activity of FeZn-PCNF and PCNF was further evaluated by the CV measurements of symmetrical batteries containing the Li₂S₆ electrolyte. As reveled by Figure 4b, FeZn-PCNF exhibits stronger current responses with three pairs of notable redox peaks, demonstrating higher catalytic activity for the bidirectional conversion of polysulfide. By virtue of the superior adsorption–catalysis effect, the FeZn-PCNF interlayer endows the Li-S batteries with improved electrochemical performances. The CV curves of FeZn-PCNF- and PCNF-based Li-S batteries at 0.1 mV s⁻¹ are exhibited in Figure 4c. The cathodic peaks C1 and C2 are attributable to the reduction of S₈ to soluble lithium polysulfides and lithium polysulfides to insoluble Li₂S/Li₂S₂, respectively. The anodic peaks A1 and A2 are assigned to the processes of Li₂S₂/Li₂S back to soluble lithium polysulfides and finally to S₈ [50,51]. in addition, the Li-S battery with the FeZn-PCNF interlayer exhibits much faster current response and smaller polarization, indicating that a more powerful catalytic effect on the reversible transformation of sulfur species. Furthermore, the CV curves of the Li-S batteries with the FeZn-PCNF and PCNF interlayers were measured at different scan rates (Figure S6), and the diffusion coefficient of lithium ion (D_Li) was calculated based on the Randles–Sevcik equation [21,52]:

$$I_p=\left(2.69\times {10}^5\right)n^{1.5}A{D}_{\mathrm{Li}}^{0.5}{v}^{0.5}{C}_{\mathrm{Li}}$$

where I_p refers to the peak current (A), n refers to the number of charge transfers in the reaction (n = 2), A refers to the electrode area (1.13 cm²), v refers to the scan rate (mV s⁻¹), and C_Li refers to the lithium ion concentration in the electrolyte. Herein, the I_p values increase with the increase in the scan rate and follow a linear relationship with v^0.5 as revealed by Figure 4d,e. Obviously, a larger slope suggests a faster Li-ion diffusion. As seen from Figure 4f, the D_Li values of the FeZn-PCNF-based Li-S battery are 2.67 × 10⁻⁷, 3.00 × 10⁻⁷, 5.11 × 10⁻⁸, and 1.60 × 10⁻⁷ cm³ s⁻¹ for the peaks A1, A2, C1, and C2, respectively, all of which are higher than those of the PCNF-based Li-S battery, demonstrating accelerated diffusion kinetics of lithium ions in the FeZn-PCNF interlayer. The EIS spectra (Figure S7) show that the charge transfer resistance of the FeZn-PCNF-based Li-S battery is much smaller than that of the PCNF-based Li-S battery, which indicates that FeZn-PCNF has higher electronic conductivity and faster interfacial catalytic activity [53].

The cycling performances of the FeZn-PCNF- and PCNF- based Li-S batteries were evaluated at 0.5 C. As displayed in Figure 5a, the initial specific capacity of the PCNF-based Li-S battery is 1021 mA h g⁻¹ and maintains at only 660 mA h g⁻¹ after 200 cycles, but the FeZn-PCNF-based Li-S battery delivers an initial specific capacity up to 1140 mA h g⁻¹ and a high retained capacity of 795 mA h⋅g⁻¹ after 200 cycles, suggesting a higher active mass utilization and a superior cycling stability. Furthermore, the microstructure and composition of the FeZn-PCNF interlayer after cycles were investigated. As observed from Figure S8a, sulfur is uniformly deposited on the surface of the nanofibrous film without large agglomerations. The EDS mappings (Figure S8b–g) show the even dispersion of C, N, Fe, Zn, and S elements. The above results confirm that the active materials would be immobilized in this conductive interlayer and be utilized during cycling. The GCD curves of both Li-S batteries (Figure S9) show two discharge plateaus and a long charge plateau. The high and low discharge plateaus correspond to the conversion from solid S₈ to soluble Li₂S_n and the conversion from Li₂S₄ to solid Li₂S₂/Li₂S, respectively, and the charge plateau corresponds to the inverse conversion process [54]. It differs in that the FeZn-PCNF-based battery exhibits a lower potential gap (∆E), implying that M-N_x active centers can propel the reaction kinetics of sulfur species. The capacity contribution at high and low discharge plateaus was further analyzed in Figure 5b,c. The generation of a high-plateau capacity (Q_H) is accompanied by the intractable shuttling of polysulfides, and the low-plateau capacity (Q_L) is restricted by the incomplete conversion due to the sluggish and nonuniform Li₂S₂/Li₂S deposition [55,56]. As revealed by Figure 5d,e, the values of Q_H and Q_L of the FeZn-PCNF-based Li-S battery are always higher than those of the PCNF-based Li-S battery along with the cycles, which fully demonstrates that the FeZn-PCNF interlayer can effectively intercept the polysulfide crossover and induce the fast and uniform deposition of solid Li₂S₂/Li₂S. The rate performances of the batteries based on two interlayers are shown in Figure 5f. The capacities of the FeZn-PCNF-based Li-S battery are 1421, 1103, 1013, and 926 mA h g⁻¹ at 0.2, 0.5, 1, and 2 C, respectively, all of which are much higher than those of the PCNF-based Li-S battery. Besides, the FeZn-PCNF-based Li-S battery still delivers a decent specific capacities based on the total mass of the cathode and the interlayer (Figure S10). Figure 5g shows the long cycling performance of the FeZn-PCNF-based battery at 1 C. After five cycles of activation, the discharge capacity is up to 949 mA h g⁻¹ and a high capacity of 618 mA h g⁻¹ is retained after 600 cycles, corresponding to a low capacity decay of only 0.06% per cycle. Based on the above discussion, the FeZn-PCNF interlayer is responsible for admirable rate and cycling performances of the Li-S batteries. The synergistic work mechanism of Fe and Zn active centers is illustrated in Figure 5h. During the discharge process, atomic Zn sites with a low electronegativity strongly capture soluble polysulfides and accelerate their reduction reaction, thus regulating the uniform deposition of insulative Li₂S. During the ensuing charge process, the Fe-N_x active sites decrease the energy barrier of the Li₂S decomposition, thus kinetically favoring the oxidation reaction from Li₂S to S. Therefore, the polysulfide shuttling is effectively mitigated and the reversible solid–liquid–solid conversion reactions are significantly boosted.

4. Conclusions

In summary, a freestanding FeZn-PCNF interlayer was successfully developed by electrospinning followed by the high-temperature pyrolysis method to physically block polysulfides and regulate sulfur chemistry. The flexible and conductive carbon skeleton serves as the interconnected “expressway network” to speed up the migration of electrons and lithium ions. With assistance from the spatial confinement of FeZn-ZIF polyhedrons and the strong anchoring of ample nitrogen atoms, Fe and Zn single atoms are uniformly doped into the carbon fibrous network and synergistically contribute to enhancing the chemical immobilization for polysulfides and propelling the redox reaction kinetics. Benefiting from these comprehensive advantages, the FeZn-PCNF-based Li-S battery demonstrates a high specific capacity of 1224 mA h g⁻¹ at 0.2 C and a high capacity retention of 908 mA h g⁻¹ even at 2 C. Meanwhile, the battery achieves an excellent cycling performance with a capacity decay of only 0.06% per cycle over 600 cycles at 1 C. This work offers a feasible and appealing route to the design of dual-metal SACs towards high-efficiency Li-S batteries and other electrochemical systems involving multi-order reactions.