Introduction

The overwhelming development in electronic device such as flexible electronic skin, portable electronic tools, and electric vehicles has tremendous impact on Today’s life [1,2,3,4,5,6]. This influence is met the demand of revolutionized the development of lithium-ion batteries [7,8,9,10]. However, the state-of-the-art lithium-ion batteries, which contained highly flammable and volatile organic liquid electrolyte, are suffered from serious safety problems (leakage, fire, weak solid electrolyte interface (SEI), and explosions), and that are difficult to meet the safety issues and specific energy requirements for future electronic devices [11,12,13,14,15,16]. Generally, the inhomogeneous SEI layer has been considered as a reason which results in the unstopped formation of lithium dendrites. Based on this, a highly uniform and stable SEI layer plays a critical role on stabilizing current flux and suppressing lithium dendrite growth, thus improving CE and long-term cycle lifespan [17,18,19]. However, traditional liquid electrolyte (LE) cannot meet the requirement as existence of undesirable side reaction with lithium metal. Therefore, replacement of the LE can be considered as the most convenient and effective method for addressing the drawbacks. As a promising substitute with LE, the usages of solid electrolyte can effectively resolve the above safety problems due to stability with electrode and exhibiting significantly suppress the lithium dendrite growth.

It is well-known that solid electrolyte is mainly divided into three categories: inorganic solid electrolyte (ISE) [20], all-solid-state polymer electrolyte (SPE) [21], and gel polymer electrolyte (GPE) [22]. ISE gives a poor interfacial compatibility with electrode and electrochemical performance at room temperature even though its deliveries a high ionic conductivity (> 10−4 S/cm), the working temperature for ISE type electrolyte often requires up to 50 °C. In case of SPE, it has optimized interfacial property, highly flexibility, and robust mechanical property compared with ISE but fails to ionic conductivity (< 10−4 S/cm) [23]. Taking full advantage of ISE and SPE, the performances of ionic conductivity and interfacial issues are improved for GPE. Nevertheless, the improved performances are sacrificed by mechanical strength. All of these restrict the practical usage in lithium-ion batteries. In this regard, a GPE with duel network (DN) can significantly address the problems of solid-state batteries. The LE trapped in the solid polymer matrix can improve electrochemical performance, ionic conductivity, and mechanical properties because beneficial from LE and framework materials.

A consecutive pore structure is favorable to ion transportation and trapping more LE, which is regarded as an ideal scaffold for immobilization of LE. Much different types of LE host materials have been reported such as most potential host materials poly (ethylene oxide) (PEO), poly (vinylidene fluoride) (PVDF), and poly (vinylidene fluoride-hexafluoropropylene) (PVDF-HFP) [24, 25]. Among these host materials, PVDF and PVDF-HFP are regarded as most suitable materials for encapsulation of LE because that it possesses the high dielectric constant, thus promoting the dissociation of lithium salts.

In this paper, we report preparation of a new polymer cross-linking PVDF nanofibrous-based quasi solid-state electrolyte (P-PVDF) to improve interfacial resistance between electrolyte and electrodes. P-PVDF is prepared by in situ polymerization of melamine and epoxy-ended amino-terminated polyoxypropylene in PVDF nanofiber framework, using LiPF6-based LE as solvent. The connected PVDF nanofiber framework can significantly improve mechanical strength comparing to single network PVDF nanofiber electrolyte, while ether-based cross-linking polymer chains endue fast ion transportation. Meanwhile, the hybrid network PVDF nanofiber electrolyte contributes to the one-off formation of a homogenous SEI to suppressing dendrite lithium. The assembled LiFePO4-based cell using P-PVDF as electrolyte exhibits significantly enhanced rate behavior with retention rate of 75% at 8 °C and cycling stability of 92% initial capacities after 200 cycles.

Experimental

Preparation of P-PVDF nanofiber polymer electrolytes

The amino-terminated polyoxypropylene (NH2-PPE, Mn = 2000, Shanghai Macklin Biochemical CO., Ltd., Shanghai) and poly (ethylene glycol) diglycidyl ether (PEDE, Mn = 400, Shanghai Macklin Biochemical CO., Ltd., Shanghai) are dissolving in ethanol. The molar radio of amino and epoxy keeps 1:2 for ensuring adequate epoxy and acquiring epoxy-ended chain extender. The reaction wormed at 50 °C for 2 h, and the product was collected by removing the solvents, named epoxy-ended PPE. The P-PVDF polymer electrolytes were prepared by in situ polymerization of precursor solution in the framework of PVDF nanofiber. The precursor consisted of 2 vol% epoxy-ended PPE and various contents of melamine (vol.melamine:vol.epoxy-ended PPE, 2%, 5%, and 10%) dissolved into 2 ml LiPF6 electrolyte (1 M LiPF6 salt in a non-aqueous mixture of EC/DEC/EMC (1:1:1 by volume), Zhangjiagang Guotai-Huarong New Chemical Materials Co., Ltd., Jiangsu). Subsequently, the precursor solution injected into a piece of PVDF nanofiber membrane (diameter = 18 mm, fabricated by electro-spinning technique as previous work [26]) and heated to 80 °C for 6 h to obtain P-PVDF. The P-PVDF polymer electrolyte contained different contents of melamine were denoted Pn-PVDF. All procedures were carried out in Ar-filled glove box (H2O < 0.1 ppm, O2 < 0.1 ppm).

Cell assembly

LiFePO4-based cathode was prepared by casting a mixture of LiFePO4 powders (Shenzhen Kejing Star Technology Co., Ltd., Shenzhen, China), carbon black (Super-P), and PVDF as binder by a weight ratio of 80:10:10 on Al foil. The electrodes were first punched into circles (D = 14 mm) with the loading content of the LiFePO4 being ~ 2.0 mg/cm and subsequently kept in a vacuum oven at 80 °C for 48 h. The lithium metal foil (battery grade, D = 16 mm, thickness = 2 mm, Shenzhen Kejing Star Technology Co., Ltd., Shenzhen, China) was used as the anode. To assemble LiFePO4-based cell using P2-PVDF as polymer electrolyte, the batteries sealed in CR2025 coin cell. The symmetric cells of double stainless steel (SS) and lithium systems were assembled by sandwiching the Pn-PVDF between two pieces of SS or lithium metal foil in a CR2025 coin cell. Lithium//Pn-PVDF//SS cells were assembled by equipping Pn-PVDF between lithium metal foil and SS in CR2025 coin cell. The whole process for the cell assembly was carried out in an Ar-filled glove box (H2O < 0.1 ppm, O2 < 0.1 ppm).

Characterization of P-PVDF polymer electrolyte

The morphology of the obtained samples was observed by scanning electron microscopy (SEM, SN-3400, Hitachi Ltd., Japan). Fourier-transform infrared spectroscopy (FT-IR) was recorded on a Nicolet 6700 spectrometer from 4000 to 500 cm−1. Tensile strength of P-PVDF was carried out on the mechanical testing system (Zwick-Roell Z005, Ulm, Germany) at strain rate of 1 mm/min and the initial gauge length was 20 mm. Electrochemical performances were investigated on electrochemical working station (CHI 660E, CH. Instruments Inc. Shanghai, China). The ionic conductivity was obtained from following equation:

$$ \sigma =L/{R}_bA $$
(1)

where the L and A are representative of the thickness and the effective area of the separator, respectively. A LAND battery cycle system (Wuhan Blue Electric Co., Ltd., Wuhan, China) was employed to investigate the battery performances (charge-discharge and cycling performance) of cells.

Results and discussion

The mechanism of cross-linked polymer is shown in Fig. 1. Firstly, the NH2-PPE reacted with PEDE to achieve two epoxy groups. Subsequently, the two epoxy groups reacted with three amino groups on the melamine through ring-opening polymerization, further growing into bulk polymer. The mechanism of polymerization is verified by FT-IR, as shown in Fig. 2a. The bands of 1654 cm−1 (N-H), 1024 cm−1 (C-O), and 885 cm−1 (epoxy group) are assigned to epoxy-ended PPE. Melamine is confirmed by the bands of 1651 cm−1 and 1554 cm−1 which are corresponding to N-H and C=N groups. The band at 885 cm−1 (epoxy group) is disappeared, while the band of 1554 cm−1 which belongs to C=N groups of melamine is detected after polymerization, suggesting cross-linking polymer successfully synthesized. Mechanical toughness not only deliveries to maintain membrane-state but also prevents physical damage during assembling process and suppresses dendrite lithium. Figure 2b presents stress-strain curves of pure PVDF and Pn-PVDFs. Nevertheless, the pure PVDF nanofiber membrane gives poor mechanical strength (0.04 Mpa). The mechanical properties enhance with increasing cross-linking intensity of polymer, and finally deliveries 0.2 Mpa for P10-PVDF. The increased mechanical strength is attributed to cross-linking polymer enhanced nanofiber.

Fig. 1
figure 1

Polymerization mechanism of the cross-linking polymer

Full size image
Fig. 2
figure 2

a FT-IR spectra of monomer and cross-linking polymer; b stress-strain curves of pure PVDF and Pn-PVDFs

Full size image

The morphology of as-prepared polymer electrolyte is displayed in Fig. 3. The morphology of pure PVDF nanofiber membrane exhibits random aggregation of fibrous structure (Fig. 3a). Much difference from loose structure of pure PVDF nanofiber membrane, the cross-linking polymer fills up the volume of PVDF nanofiber and further connects each other into a compact structure (Fig. 3b). From EDXS elemental mapping, C and F signals recording in P-PVDF which are characteristic elemental of PVDF are maintained nanofibrous structure, the N signal also consistently distributes in P-PVDF which are stemmed from the group of –NH group in cross-linking polymer, suggesting existence of cross-linking polymer.

Fig. 3
figure 3

Morphology of pure PVDF and P-PVDF

Full size image

The plots of log σ vs. T−1 for as-prepared P-PVDF polymer electrolyte present in Fig. 4a. The ionic conductivity of pure PVDF is as high as P2-PVDF, reaching value of 1.35 mS/cm at 25 °C. That value is much higher than LE which only gives 0.95 at 25 °C (see Fig. S1a). However, the contents of melamine further increasing, the ionic conductivity dramatically reduced to 0.74 mS/cm (P5-PVDF) and 0.53 mS/cm (P10-PVDF), indicating that the suitable immobilization of melamine for P-PVDF polymer electrolyte is about 2%. The decreased ionic conductivity is attributed to cross-linking intensity. In high cross-linking intensity, the mobility of polymer chains are impeded; as a result, the movement of ion is hindered by the ponderous polymer chains. The activation energy (Ea) refers to ionic transportation. The lower value of Ea means rapid ionic transportation. The Ea is calculated from Vogel-Tamman-Fulcher (VTF) empirical equation [27]:

$$ \sigma ={\sigma}_0{T}^{-1/2}\exp \left(-\frac{E_a}{R\left(T-{T}_0\right)}\right) $$
(2)

where Ea is the activation energy, σ0, T0, and R are constant factors for the pre-exponential factor, parameter corrected to the gas transition temperature and ideal gas constant, respectively. The results are similar with ionic conductivity. The P2-PVDF polymer electrolyte exhibits minimum Ea (~ 2.3 × 10−2 eV) that much lower than other polymer electrolytes (~ 3.2 × 10−2 eV for pure PVDF, 4.3 × 10−2 eV for LE, 4.8 × 10−2 eV for P5-PVDF, 5.5 × 10−2 eV for P10-PVDF). This means the ion could easily transmitted in the P2-PVDF electrolyte. The remarkable Ea are originated from ether-dominated cross-linking polymer. The ether group is intimate with LE, which means more LE are trapped in the ether-dominated cross-linking polymer network. Therefore, the cross-linking polymer immobilization of LE is acted as “lithium-ion tank” which storages into the PVDF nanofiber framework. For the lithium-ion batteries, we prefer to cation (lithium ion) transportation in the batteries system, rather than anion transmitting due to mechanism of lithium-ion batteries. Based on this, lithium-ion transference number (t+) plays a critical role to evaluate whether the electrolyte could apply in lithium-ion batteries. As seen in Fig. 4b. Comparing with LE (0.32, see Fig. S1b), P5-PVDF (0.24) and P10-PVDF (0.19), the pure PVDF and P2-PVDF electrolyte gives relatively high t+, which achieves as high as 0.36 and 0.48, respectively. It depends on synergistic effect of polymer chains to migrate lithium ion. The –F group in the PVDF could bond with lithium ion, thus the lithium ion could migrate with PVDF polymer chains. In the P-PVDF electrolyte, ether-dominated cross-linking polymer chains have noticeable compatibility with lithium ion; hence, the lithium-ion movement could be significantly improved by double transport network (PVDF and ether-rich cross-linking polymer). In case of high melamine, the ether-rich polymer chains are highly cross-linked by melamine, as consequence of poor chain mobility. The ponderous polymer chain significantly impeded lithium-ion transportation in the electrolyte, thus resulting in low t+. These illustrate the suitable cross-linking intensity should keep the melamine content at 2%.

Fig. 4
figure 4

a Arrhenius plots of a polymer electrolyte based on the pure PVDF and Pn-PVDF; b chronoamperometry profiles of the pure PVDF and Pn-PVDF at 25 °C, using a step potential of 10 mV

Full size image

The lithium plating/stripping cycling for the cell based on P2-PVDF polymer electrolyte was tested by using lithium metal symmetric cell. The symmetric lithium cell based on P2-PVDF as electrolyte gives a low voltage hysteresis ~ 86 mV at 1 mA/cm2 (Fig. 5a). It is noteworthy that no evidence voltage drop have been occurred even the cycling after 300 h, as indication of good performance on controlling dendrite lithium growth. However, in the case of LE-based symmetric lithium cell, the voltage hysteresis is as high as ~ 190 mV at initial stage much higher than that of P2-PVDF (Fig. S2a). The voltage hysteresis further increases as the cyclic number gaining and finally stops at 400 mV, presenting poor stability of plating/stripping performance. The poor stability indicates rapid lithium dendrite growth on the surface of lithium metal. The improved dendrite resistance for P2-PVDF is attributed to formation of homogeneous SEI layer, which is further confirmed by SEM and XPS. In the Fig. 5b, the morphology of the SEI layer formed by P2-PVDF presents a smooth and flat without any defects, suggesting no obvious rigid dendrite lithium forming. Much contrast with SEI layer formed by P2-PVDF, the surface of lithium metal becomes much rough and irregularity with visible of agglomerate dendrites for the LE-based cell (as seen in Fig. S2b). The results are stemmed from superior compatibility between P2-PVDF and lithium electrode. It is further verified by electrochemical impedance spectroscopy (EIS, as shown in Fig. S3). The Nyquist plot for P2-PVDF exhibits a typical semicircle. The high-frequency semicircles are assigned of the charge transfer resistance on the electrode/electrolyte interfaces. The initial interfacial resistance of the lithium//P2-PVDF//lithium is ~ 165 Ω. The resistances are positive correlation with the storage time for the cells, owing to active reactions between the interface of bared lithium and electrolyte. The dramatically increased resistances mean inevitable and overwhelming side reactions is occurring. For the cell based on P2-PVDF as electrolyte, the resistance shows a negligible change (~ 75 Ω increased) even when the storage time lasted for 10 days, as an indication of superior compatibility between P2-PVDF and lithium metal [28,29,30]. For confirming the chemical composition of SEI, the chemical structure of SEI layer was confirmed by XPS. The high-resolution C 1s spectra of SEI formed by P2-PVDF divided into four peaks which locates at 284.3, 286.7, 288.6, and 290.5 eV, corresponding to C-C, C-O, C=O, and RO-COO, respectively (in Fig. 5c). The existence of RO-COO reflects decomposition of organic solvents of liquid electrolyte. A low concentration of RO-COO can be found in the C 1s spectra, indicating formation of ultra-stable SEI layer. Three peaks around 684.5, 686.3, and 688.7 eV can be observed from F 1s spectra (Fig. 5d), reflecting the formation of Li-F LixPOYFZ, and –CF3, respectively. The concentration of LixPOYFZ and –CF3 are much lower than that of LiF. The SEI layer has more LiF, considering more efficiently suppress the growth of dendrite lithium. In case of the LE-based cell, the concentration of C-O group in the C 1s is lower than that of the C-C group, as presented in Fig. S2c. It indicates the compositions of SEI for the LE-formed are mainly consist of inorganic lithium salts, which is reported in our previous work [31]. It is also can be seen from F 1s XPS spectra for SEI layer formed by liquid electrolyte. A visible concentration of lithium salts (686.5 eV, in Fig. S2d) can be observed, further confirming the inorganic salts-dominated SEI layer. Moreover, the peak intensity of –CF3 is much lower than that of LiF in the SEI layer formed by the P2-PVDF. The inorganic salts-based SEI layer is very fragile and unstable, resulting in ongoing consuming lithium ion to “repairing” SEI layer.

Fig. 5
figure 5

a Long-term lithium plating/stripping cycling of the cell based on P2-PVDF; b morphology of SEI layer formed by the cell using P2-PVDF as electrolyte; XPS spectra of SEI layer c C 1s and (d) F 1s

Full size image

To evaluate the practical usage of P2-PVDF electrolyte, the full cell based on LiFePO4//P2-PVDF//lithium are assembled and investigated by series of electrochemical performance, as shown in Fig. 6. Figure 6a presents the initial charge/discharge curves of the cell based on LE and P2-PVDF with potential from 2.75 to 4.2 V. A 0.1 C (25 mA/g) current density carried out for the cell. A couple of electrochemical platform for LiFePO4 (3.45 V for charge, 3.35 V for discharge) relates to oxidation and reduction potential. The specific charge/discharge capacities of P2-PVDF based cell are 159.5 and 158.6 mAh/g slightly higher than the cell based on LE which gives 158 and 157 mAh/g; respectively. The rate behavior is essential for rapid charge/discharge performance. Figure 6 b profiles the discharge capacity of the cell based on P2-PVDF at different current density (0.1 C, 0.5 C, 1 C, 3 C, 5 C, and 8 C). In case of relative low current density (0.1 C), the discharge capacity is as high as ~ 160 mAh/g for both cells based on LE (see Fig. S3a) and P2-PVDF (see Fig. 6b). Increasing current density, the discharge capacity drops for the both cell. Obviously, 75% of initial capacity is still maintained for the cell equipping P2-PVDF as electrolyte after running at 8  C, the value is much higher than that of the cell assembled LE which only keeps 61% of initial capacity (Fig. S4a). The obtained results are almost higher than that of most recently reported polymer electrolyte (Table S1), as indication of a strong potential for applying in high-performance rechargeable lithium-ion batteries. The superior rate behavior is owed to contribution of high ionic conductivity and remarkable t+. Consideration of remarkable lithium plating/stripping loop, the as-assembled full cell also exhibits noticeable cyclic lifespan. In Fig. 6c, no evidence capacity fading have been observed after the cell performed 200 cycles, the specific capacity is as high as ~ 139 mAh/g, meaning that 92% of initial capacity is still retained. Moreover, the Coulombic efficiency for each cycle are high than 98%. Much contrast from the cell using P2-PVDF as electrolyte, it is only 80% of initial capacity maintained after 200 cycles for the LE-based cell, and the Coulobmic efficiency drops to ~ 95% with cyclic test proceeding (Fig. S4b). The remarkable electrochemical stability is attributed to robust and inhomogeneous SEI layer. For further understanding the excellent cyclic stability of the cell based on P2-PVDF, the difference of resistance tested by EIS measurement was used to evaluate the interfacial behavior of lithium metal cells (as displayed in Fig. 6d). The fist intercept of the semicircle on the EIS plot represents the bulk resistance of cell (Rb); diameter of the semicircle gives the value of charge transfer resistance Rct. As seen from Fig. 6d, Rct of the cell equipping P2-PVDF as electrolyte, corresponding to migration of lithium ion between the electrode and electrolyte interface, is only 74.3 Ω. The resistance gains to 136.7 Ω, as the cyclic numbers ended at 200 cycles. The gained resistance owes to formation of SEI layer and inevitable irreversible side reaction between bare lithium and electrolyte. The difference between initial and cycled ones determined the level of polarization [32,33,34]. Therefore, the resistance of P2-PVDF compared to initial one only raises 62.4 Ω which is much lower than that of LE (183 Ω, see Fig. S4c). This small resistance indicates the cell using P2-PVDF as electrolyte keeps a low polarization state which benefits from formation of highly homogenous SEI layer, rather than that of sustaining consumption of lithium metal to maintain intact SEI layer [35, 36].

Fig. 6
figure 6

The electrochemical performance of the cell a initial charge-discharge capacities of LE and P2-PVDF, b rate behavior, c cycling performance at 1 °C, and d the resistance gap between initial and cycled

Full size image

Conclusions

In this paper, P-PVDF is prepared via one-pot synthesis strategy based on in situ polymerization of melamine and epoxy-ended amino-terminated polyoxypropylene with LiPF6-based liquid electrolyte in framework of PVDF nanofiber. The polymer cross-linking PVDF nanofiber not only improves mechanical strength for polymer electrolyte but also contributes to the one-off formation of a homogenous SEI layer to suppressing dendrite lithium. The ether-dominated polymer chains provide ionic transportation channel induced high ionic conductivity (1.35 mS/cm). As consequence, the as-assembled cell presents superior electrochemical performance. In comparison with LE, the P-PVDF polymer electrolyte exhibits significantly enhanced rate behavior with retention rate of 75% (61% for the liquid electrolyte) at 8 C, cycling stability of 92% initial capacities (80% for liquid electrolyte) after 200 cycles.