The Effects of Electrospinning Structure on the Ion Conductivity of PEO-Based Polymer Solid-State Electrolytes

Abstract

To overcome the safety hazard of the liquid electrolytes used in traditional lithium batteries, solid electrolytes have drawn more attention because of their advantages such as non-volatility, easy processing, good mechanical properties, and stability. In this paper, sodium alginate (SA) nanofiber membranes were used as the backbone of PEO-based solid electrolytes. SA nanofiber membranes were prepared by electrospinning with assistance from PEO and cross-linked with calcium ions to construct a nanofiber network skeleton, which provided a guarantee for the stability of the subsequent electrolyte preparation process. The effects of spinning conditions and crosslinking time on the structure and performances of the nanofiber membranes were investigated. Meanwhile, the relationship between the skeleton of nanofiber membranes cross-linked with calcium ions and ion conductivity was investigated. The optimal parameters of the electrospinning process including concentration, voltage, distance, and SA content were discussed, and the fiber diameter and its distribution were analyzed. Furthermore, Fourier transform infrared (FTIR) spectrometer, thermal gravimetric analyzer analysis (TGA), X-ray diffraction (XRD), and energy dispersive spectrometer (EDS) maps were used to characterize the nanofiber membranes and electrolytes. The results showed that the thermal performance of cross-linked nanofiber membranes improved and the crystallinity of the PEO matrix decreased. The ion conductivity of the electrolytes was characterized by electrochemical impedance spectroscopy (EIS) testing, and the results showed that the assembled lithium symmetric battery had a good ion conductivity of 6.82 × 10⁻⁵ S/cm at 30 °C.

1. Introduction

Lithium batteries are mainly used in digital products in traditional fields, but with the development of the global new energy industry, the demand in this field has been relatively saturated [1]. In emerging fields, lithium batteries are mainly used in power batteries and energy storage. However, the organic liquid electrolyte used in traditional lithium batteries has many potential safety hazards, such as poor heat dissipation, easy leakage, and uncontrolled lithium dendrite growth. These issues can easily lead to safety issues such as battery short circuits, fires, and even explosions, thereby limiting the practical application of lithium batteries [2]. Compared with liquid electrolytes, solid electrolytes not only have excellent mechanical properties and low flammability but also own obvious advantages in thermal stability and electrochemical stability, which greatly improve the safety performance of lithium batteries [3]. The solid electrolyte is hopeful to replace the liquid electrolyte, which can overcome the lithium dendrite growth problem and fundamentally solve the safety problem. In addition, during the assembly process of a solid-state battery, the solid electrolyte can simultaneously replace the diaphragm and electrolyte, which can simplify the internal structure of the battery and the assembly process while reducing the cost. In the future, the weight and volume of the battery pack can be reduced through reasonable design, to achieve the purpose of improving the energy density of the battery pack, making it more durable. Therefore, all-solid-state lithium batteries are the most promising next-generation rechargeable energy storage devices for development [4].

At present, solid electrolytes mainly include inorganic solid electrolytes (ISEs), polymer solid electrolytes (PSEs), and composite solid electrolytes (CSEs) [5]. Among them, ISEs have high ionic conductivity and lithium–ion migration number, wide electrochemical window, and excellent electrochemical stability, but poor interface compatibility with electrodes and brittleness [6]. PSEs have good interfacial compatibility and flexibility, but poor mechanical strength and low ionic conductivity. CSEs combine the advantages of ISEs and PSEs and have high ionic conductivity and excellent mechanical properties [7]. Polyethylene oxide (PEO) is the earliest and most widely studied solid polymer electrolyte material with strong lithium salt solubility, and also the most prevalent solid electrolyte material [8]. This is because it can form complexes with various alkali metal cations and has a high dielectric constant. In addition, it can effectively dissociate lithium salts and has good ion conductivity at high temperatures. However, PEO mainly relies on amorphous regions for lithium conduction. Its poor mechanical strength and low ionic conductivity at room temperature limit the further development of PEO-based solid electrolytes [9]. Therefore, the research on PEO mainly focuses on the improvement of ionic conductivity at room temperature, mainly through physical or chemical modification to lower the crystallinity of PEO, so that more amorphous PEO can participate in ion transport at room temperature [10]. The use of blends, copolymers, and comb-branched polymers is a typical method for increasing ion conductivity, which aims to improve ion conductivity by reducing the crystallinity of polymer electrolytes [11]. In PEO-based polymer solid electrolytes, lithium ions undergo complex reactions with oxygen within or between PEO molecular chains. Accompanied by the formation and cleavage of lithium–oxygen (Li-O) bonds, lithium ions undergo jumps between different coordination sites through coordination and uncoordination, while the peristalsis of polymer segments within the free volume promotes their directional migration [12].

In recent years, the constructional design of solid electrolytes has developed into a research highlight. Various solid electrolytes were designed and prepared with excellent comprehensive performance, promoting the development of all-solid-state batteries [13]. Liu’s team [14] infiltrated nylon through the sintered gel to synthesize Li_0.33La_0.557TiO₃ (LLTO) framework with an interlocking porous structure. Then, polyethylene oxide (PEO) was added to the pores to generate a PEO–LLTO framework solid electrolyte (PLLF electrolyte) with vertical bicontinuous phases. PLLF electrolyte has excellent ionic conductivity at 25 °C, reaching 2.04 × 10⁻⁴ S cm⁻¹. The vertical continuous framework and closed PEO, as effective Li⁺ transmission highways, significantly improve the ion conduction of electrolytes. In addition, the interconnected networks give PLLF electrolytes outstanding structural stability. Therefore, the battery exhibits extraordinary cycling stability. Bai’s group [15] presents a polymer–ceramic hybrid electrolyte (polyethylene oxide (PEO)/polyvinylidene fluoride (PVDF)/Li_6.4La₃Zr_1.4Ta_0.6O₁₂ (LLZTO)) is designed and modified with trace amounts of liquid electrolytes. The addition of PVDF not only reduces the crystallinity of PEO polymer but also enhances the affinity between the liquid electrolyte and the composite electrolyte. At the same time, adding a small amount of liquid electrolyte at the interface of the composite electrolyte/electrode stabilizes the interface layer after low current charging and discharging, avoids contact loss during the initial state and cycling process, and reduces interface resistance. Therefore, the assembled LiFePO₄/Li batteries display a high discharge capacity of 160.1 mA h g⁻¹ and the long cycle life of Li symmetric batteries is not short at around 1000 h. The composite solid electrolyte provides an effective and feasible method for solving interfacial problems and developing high-performance solid lithium metal batteries. Zhang’s team [16] prepared a composite solid electrolyte, which combined a PAN/LLZTO fiber network prepared by PEO electrospinning with PEO polymer. The fiber/ceramic composite material network with a three-dimensional structure was uniformly distributed in the PEO polymer matrix, which significantly reduced the crystallinity of the PEO polymer, thus successfully constructing a continuous Li⁺ transport channel, and improving the mechanical strength of the PEO matrix. The later put-together LiFePO₄/PPL/Li batteries represent outstanding electrochemical performance at both room temperature and high temperature. However, the influence of electrospinning structure on ionic conductivity is still unclear, which restricts the application of electrospinning in polymer-based solid electrolytes.

In this paper, inspired by the three-dimensional support network framework and various cross-linking modification strategies, SA was introduced to a nanofiber membrane by electrospinning with assistance from PEO and then cross-linked with calcium ions, which was used as the backbone of PEO-based solid electrolytes. The relationship between the skeleton of nanofiber membranes cross-linked with calcium ions and ion conductivity was investigated. The PEO matrix in the fiber network can improve the ionic conductivity and electrochemical stability of the electrolytes. The nanofiber network structure prepared from alginate serves as the support skeleton for the solid electrolytes, which can maintain the morphology stability of the PEO matrix at high temperatures and effectively reduce the safety hazards caused by the flammability of PEO itself. The excellent performance of SA perfectly meets the demand for improving the safety performance of PEO-based solid-state electrolytes.

2. Experimental Methods

2.1. Materials

PEO (M_w = 300,000 and 600,000), lithium bis(tri-fluoromethane) sulfonimide (LiTFSI, 99.5%), and anhydrous Acetonitrile (99.8%) were provided by Aladdin in Shanghai, China. SA (20 °C, 1% discharged water solution) was provided by Haizhilin in Qingdao, China. Anhydrous calcium chloride (CaCl₂, 96%), 2, 2′-Azobis (2-methyl propionitrile) (AIBN, 98%), and PEGdMA (750 Da) were purchased from Macklin in Shanghai, China. Li_6.4La₃Zr_1.4Ta_0.6O₁₂ (LLZTO, 5 μm) was provided by Kejing in Hefei, China. PEO and LiTFSI were dried in a vacuum at 45 °C and 100 °C, respectively, for 24 h before use.

2.2. Preparation of PEO/SA Composite Membranes and Cross-Linked Membranes

The spinning solution was prepared by dissolving SA and PEO (M_w = 600,000) in deionized water, stirred mechanically for 4 h to be mixed evenly, and vacuum defoamed. Then, 15 mL of uniform and steady solution was sucked slowly into a syringe with a metal needle tip (length of 40 mm, inner diameter of 0.4 mm). The spinning solution with different concentrations (2 wt%, 3 wt%, 4 wt%, and 5 wt%) and different compositions (the ratio of SA to PEO 10:90, 20:80, 30:70, and 40:60) was spun under various voltage (20 kV, 25 kV, 30 kV, and 35 kV) with the tip-to-collector distance (TCD) of 10 cm, 15 cm, 20 cm, and 25 cm, respectively. All samples were collected under the conditions of 27 ± 2 °C temperature and 40 ± 1% relative humidity. The nanofiber membranes were immersed into 4 wt% calcium chloride solution (ethanol: deionized water = 1:7) to crosslink and then washed in deionized water to remove the excess solution. Then, it was dried to obtain a cross-linked membrane.

2.3. Preparation of Solid Polymer Electrolytes (SPE)

First, 1.0 g PEO (M_w = 300,000) and 0.5 g LiTFSI (99.5%) were dissolved in 10 mL anhydrous acetonitrile under stirring. Then, 0.5 g PEGdMA, 0.53 g LLZTO, and 0.048 g AIBN were added into the above solution, and stirring magnetically was continued for 24 h. The solution was coated on the cross-linked membrane and vacuum dried at 80 °C for 24 h to obtain PEO/SA solid polymer electrolyte (PEO/SA-SPE).

2.4. Characterization

The structural and morphology information of the obtained morphology of nanofiber membranes were collected using a Hitachi regulus 8100 scanning electron microscope (SEM, Tokyo, Japan). The elemental analysis of electrolytes was performed by energy dispersive spectroscopy (EDS, X-max). By scanning electron microscopy (SEM) images (5000×), the average fiber diameter of the nanofibers was determined by measuring the diameters of 100 different nanofibers at different points. The following equation calculates the values of the coefficient of variation (CV):

$$\mathrm{CV} =\frac{\mathrm{standard} \mathrm{deviation}}{\mathrm{average} \mathrm{value}}\times 100\%$$

The Fourier transform infrared (FTIR, Chelmsford, Massachusetts, USA) spectra were collected from 400 cm⁻¹ to 4000 cm⁻¹ using a Nicolet iS50 spectrometer with ATR mode. The thermal properties of nanofiber membranes and cross-linked nanofiber membranes were characterized using a TGA5500 thermogravimetric analyzer with a heating rate of 20 °C min⁻¹ from 30 °C to 600 °C in a nitrogen atmosphere. The electrochemical impedance spectroscopy (EIS) was studied using a CHI660E electrochemical workstation with a frequency range of 0.1 Hz–1.0 M Hz from 35 to 80 °C. The battery measured ion conductivity by clamping the electrolyte between two lithium sheets. The calculation equation for ion conductivity (σ, S cm⁻¹) is as follows:

$$\mathsf{\sigma } =\frac{\mathrm{L}}{\mathrm{SR}}$$

where σ is the ionic conductivity, L is the thickness of the electrolytes, S is the contact area of lithium flake and electrolytes, and R is the bulk resistance of electrolytes.

3. Results and Discussion

3.1. Effects of Spinning Solution Concentrations

The spinnability of a spinning dope mainly depends on the concentration of the polymer when the electrospinning is carried out at room temperature. Therefore, in this study, we first discussed the influence of polymer concentrations on the fabrication of fiber membranes. The spinning voltage, receiving distance, and mass ratio of SA to PEO solution were fixed at 30 KV, 15 cm, and 10:90, respectively. As can be seen from Figure 1, continuous fibers were all successfully obtained at the concentrations of 2, 3, 4, and 5 wt%. The scanning electron microscope images shown in Figure 1 demonstrate that whereas beads-on-string formed at the concentration of 2 wt% (Figure 1a), essentially no beads-on-string were observed on the fibers spun from the solutions with the higher mass concentrations on the solution of 3–4 wt% (Figure 1b,c). In 4 wt% solution, the as-collected sample showed a good morphology nanofiber membrane with an average fiber diameter of 144 ± 34 nm. This was because when low concentration and viscosity of the electrospinning solution were used to spin fiber membranes, the molecular chains in the solutions were not entangled or not entangled enough, resulting in the aggregation of molecular chains and ultimately forming beads-on-string [17]. The viscosity of spinning solutions with different concentrations is shown in Figure 2. The variation of solution viscosity with solution concentration mainly included two parts: on the one hand, when the solution concentration was less than the critical entanglement concentration, the molecular chains in the solution could not form effective entanglement, making it difficult to form effective continuous fibers; on the other hand, when the concentration of the solution was greater than the critical entanglement concentration, the entanglement between molecular chains was increased, providing conditions of the continuous forming of fibers. It could be seen from Figure 2 that with the increase of solution concentration, the solution viscosity increased, but there was a boundary point for this increasing trend, which corresponds to the critical entanglement concentration that could realize electrospinning forming. When the concentration and viscosity of the solution exceeded a certain critical value, the degree of entanglement between molecular chains increased, resulting in a continuous fiber structure [18]. However, solutions with polymer concentrations exceeding 4 wt% were too viscous, resulting in the formation of droplets due to the high surface tension of the solution (Figure 1d) [19].

3.2. Effects of Spinning Voltage

Under the action of the electric field, electrospinning droplets deform into Taylor cones and then spray to form fibers under critical voltage. To explore the effects of voltage on the formation of fibers, we fixed the spinning dope concentration, the receiving distance, and the mass ratio of SA to PEO solution at 4 wt%, 15 cm, and 10:90, respectively. Figure 3 shows the SEM images of the fiber spun from SA solutions at various voltages. At a voltage of 20 kV (Figure 3a), the surface tension of the solution was greater than the electrostatic repulsion due to the low voltage, resulting in the solvent being unable to evaporate and ultimately unable to solidify into fibers [20]. As the voltage increased, the fiber could solidify and shape gradually. However, the diameter of the fibers increased as the increase of the electric field. When the voltage was set as 25 kV (Figure 3b), the average diameter of the fibers was 114 ± 64 nm, while the average diameter of the fibers increased to 144 ± 34 nm at the voltage of 30 KV. This is mainly because as the voltage increased, the tensile force and acceleration of the electric field on the jet increased, resulting in a shorter time for secondary stretching, so a larger fiber diameter appeared [21]. When the voltage reached 35 kV (Figure 3c), the jet velocity accelerated by the high voltage, and the solvent evaporation time shortened, resulting in irregular fibers formed with a large number of beads-on-string and large areas of adhesion between the fibers.

3.3. Effects of TCD

The distance between the needle and collector directly affects the molding and quality of fibers. The influence of the TCD was discussed for the electrospinning of SA from aqueous solutions in Figure 4. In this part, the spinning dope concentration, spinning voltage, and the mass ratio of SA to PEO solution were fixed at 4 wt%, 30 KV, and 10:90, respectively. The small TCD (10 cm) would have a short jet, leading to the incomplete evaporation of the solvent, thus the collected fibers were severely damaged and unable to form an intact fiber membrane [22]. As TCD increased from 15 to 20 cm, the fiber diameter decreased slightly, and the mean fiber diameter decreased from 144 ± 34 nm (Figure 1c) to 137 ± 47 nm (Figure 4). The reason may be that as the TCD increased, the jet was fully stretched and the solvent was fully evaporated, resulting in a decrease in fiber diameter [23]. However, when the TCD was 25 cm, the excessively long TCD reduced the electric field intensity, which was not conducive to the volatilization of the solvent and caused the inability to solidify into fibers [24].

3.4. Effects of SA and PEO Ratio

To better understand the influence of SA and PEO ratio on the fiber membrane formation, we further studied the fabrication of fiber membranes under different SA to PEO ratios (the spinning dope concentration, voltage, and TCD were 4 wt%, 30 KV, and 15 cm). As shown in Figure 1c, when the mass ratio of SA:PEO was 10:90, the average fiber diameter was 144 ± 34 nm (diameter distribution, 100–280 nm). No matter if the mass ratio of SA and PEO changed to 20:80 or 30:70, well-distributed nanofibers could be obtained. As the mass ratio of SA and PEO changed to 20:80, the nanofibers with a mean diameter of 122 ± 43 nm (diameter distribution of 80–180 nm) were obtained (Figure 5a). When the mass ratio of SA to PEO was 30:70, the nanofibers with an average diameter of 123 ± 53 nm and diameter distribution of 60–300 nm were obtained (Figure 5b). By improving the ratio of SA content (SA:PEO, 40:60), the fiber diameter was shortened to 103 ± 38 nm (Figure 5c, diameter distribution, 60–160 nm) and showed adhesion and beads-on-string.

The molecular chains of SA closely overlap with each other. Due to the rigidity and extended chain conformation of SA, it was difficult for SA macromolecules to form an entanglement network, so it was difficult to electrospin from the pure SA solution [25]. Adding proper polymers with high molecular weight could greatly improve the electrospinning performance of SA. Some polymers with high molecular weight have long and flexible molecular chains, making it easy to form physical networks even at low concentrations. When SA joined, it would be wrapped in its large physical network, forming an effective entanglement network. In addition, the CV value of the fibers indicates the distribution uniformity of the fiber diameter. The smaller the CV value, the more uniform the fiber diameter, as shown in Figure 6. When the concentration of the solution was 4 wt%, the voltage was 30 kV, the TCD was 15 cm, and the mass ratio of SA and PEO was 10:90, the CV value was the smallest, indicating that the fiber diameter distribution was the most uniform.

3.5. Characterizations of Cross-Linked Membranes and PEO/SA-SPE

The morphology of the nanofiber membranes cross-linked at different times was characterized, as shown in Figure 7. As the crosslinking time increased, the degree of Ca²⁺ crosslink gradually increased; at the same time, the swelling of fibers caused the pores between fibers to become smaller. The existence of fiber pores was beneficial for the subsequent preparation of electrolytes. From the figure, it could be seen that when the crosslinking time did not exceed 7 min, the fiber pores remained good. The surface morphology of the solid electrolytes was observed by scanning electron microscope. The dispersion of ceramic particles was a key factor influencing the function of electrolytes. When LLZTO ceramic particles were led into the PEO matrix, the electrolyte film had a positive morphology, and the ceramic particles were regularly distributed in the PEO matrix. It has been known that EDS mapping is a good characterization method for detecting the distribution status of elements in samples [26]. The surface EDS mapping (Figure 8) of the electrolyte showed a uniform distribution of elements C, O, La, Zr, and Ta, further indicating the regular dispersion of ceramic particles in the PEO matrix, which helped to avoid micro short circuits within the battery.

Figure 9a shows the FTIR spectra of the pure SA, PEO membrane, PEO/SA composite membrane, cross-linked membrane, and PEO/SA-SPE in the wavelength ranges of 4000–400 cm⁻¹, respectively. The absorption bands of -CH appeared at 2878 cm⁻¹ in the PEO membrane, the absorption bands of C-O-C appeared at 1095 cm⁻¹, while there were absorption bands at 1594 cm⁻¹ and 1405 cm⁻¹ in SA. Due to the addition of LiTFSI, PEGdMA, LLZTO, and AIBN in the final synthesis stage of the electrolyte, these substances would generate new peaks in the infrared spectrum. However, because of their low content, their peaks were relatively small, resulting in obvious impurity peaks near 1095 cm⁻¹. We added crosslinking agents and initiators during the production of the electrolyte, and their vibrational modes interacted with the vibrational modes we measured, resulting in the broadening of the peak. While the original characteristic absorption peak existed in the cross-linked nanofiber membrane, a broad peak appeared at 3350 cm⁻¹, representing the stretching vibration of the O-H bond. This broad peak could also be seen when processing the SA curve alone, but its intensity was weak, so it was masked after normalization with other curves. Then, under the influence of cross-linking, the original O-H stretching vibration peak on the molecular chain was enhanced by hydrogen bonding and appeared in the figure. PEO/SA composite membrane and cross-linked membrane contained all the typical characteristic peaks mentioned above, indicating the successful preparation of the samples.

It was necessary to study the thermal decomposition behavior of solid electrolytes to cope with extreme situations in use and prevent safety accidents such as fires. The thermostability was studied by thermogravimetric analysis (TGA), shown in Figure 9b. The TGA curves of the PEO membrane, PEO/SA composite membrane, cross-linked membrane, and PEO/SA-SPE mainly exhibited two thermal weight loss platforms. The thermal decomposition temperatures of the nanofiber membrane and electrolytes were about 240 °C caused by the breaking of glycosidic bonds in seaweed polysaccharides, and the other temperature 400 °C was caused by the decomposition of PEO. The cross-linked membrane would be accompanied by a small amount of carbonization when the glycosidic bond broke. The addition of calcium ions would generate metal oxides or metal carbonates and other substances in the combustion process. The final residual rate of the sample increased with the addition of alginate.

The lithium–ion conductivity of PEO was closely related to its amorphous region. Therefore, X-ray diffraction analysis was conducted to characterize the crystallinity of the solid electrolytes. As shown in Figure 9d, there were two typical crystallization peaks in the PEO membrane at 19.1° and 23.2°. In contrast, the SA nanofiber membrane exhibited the same type of crystallization peak at these two positions, and there were no other significant peaks, indicating that the addition of SA would not produce a new crystalline state. At the same time, the crystallization peak of PEO significantly weakens in the SA nanofiber membrane curve, reflecting a decrease in the crystallinity. The spectra of the cross-linked nanofiber membrane showed a halo peak representing an amorphous form, indicating that the crystallinity decreased significantly after cross-linking. The reduction of crystallinity was mainly due to the cross-linking network structure formed by hydrogen bonding between SA and PEO, as well as the physical network structure provided by the nanofiber membrane distributed in the matrix, which disrupted the regularity of PEO and gave it more amorphous regions with strong chain segment movement ability. In the final preparation process of the electrolyte, the electrolyte solution was coated on the nanofiber membrane. Due to the addition of LLZTO ceramic powder to the solution, the crystallinity of the PEO matrix could be reduced [27].

To visually characterize the changes in ion conductivity of PEO-based solid electrolytes after adding nanofiber membranes with different crosslinking times, we assembled a symmetrical battery with the electrolyte sandwiched between lithium sheets for electrochemical impedance spectroscopy testing [28], as shown in Figure 10a. According to the ion conductivity formula, the ion conductivity of the electrolyte at different crosslinking times was calculated (Figure 10b). Different crosslinking times would have a certain impact on the ion conductivity. When the crosslinking time was 5 min, the ion conductivity was the highest at 6.82 × 10⁻⁵ S/cm.

4. Conclusions

In summary, by adjusting the parameters such as spinning dope concentration, voltage, distance, and SA content of the electrospinning process, the nanofiber membranes with good and uniform morphology were obtained, which was conducive to the improvement of electrolyte stability. The results showed that the viscosity of the solution increased and the fiber diameter gradually increased as the concentration of the spinning solution increased. Simultaneously, excessive or insufficient voltage and TCD were not conducive to fiber formation. The average diameter of the fiber showed a trend of first increasing and then decreasing as the voltage increased. As TCD increased, the average fiber diameter gradually decreased. The optimal process parameters of the electrostatic spinning of SA nanofiber membranes were as follows: spinning dope concentration of 4 wt%, a voltage of 30 kV, TCD of 15 cm, and a mass ratio of SA and PEO of 10:90. The microstructure of the nanofiber membranes after crosslinking was influenced greatly by the crosslinking time. The fibers swelled significantly and the pores inside the membranes became smaller, and the degree of Ca²⁺ crosslinking increased as the crosslinking time increased. Meanwhile, after crosslinking with calcium ions, the crystallinity of the PEO matrix significantly decreased. The addition of Ca²⁺ led to the formation of metal oxides or metal carbonates on the electrolyte’s surface during thermal decomposition, which increased the residual carbon rate of the electrolyte. A composite PEO/SA-SPE for safe lithium batteries was successfully prepared in this work and the ionic conductivity of the SPE was 6.82 × 10⁻⁵ S/cm at 30 °C, showing excellent lithium–ion transport capacity.