1 Introduction

Lithium-ion batteries (LIBs) are widely used as energy storage devices due to their high energy density. Meanwhile, as its energy density continues to increase, its safety is facing challenges due to the use of flammable liquid ester electrolytes [1, 2]. Among many strategies for improving the safety of LIBs, the use of solid-state electrolytes is intensively attempted due to their non-flammability [3,4,5]. In particular, polymer-based solid-state electrolytes have attracted significant attention due to their low cost, good processing performance and interface contact, among which polyoxyethylene (PEO) polymer electrolyte is the most widely studied candidate [5,6,7,8,9]. The oxygen atoms in the ether oxygen segment of PEO can be complexed with Li+, so Li+ will undergo complexation–decomplexation–complexation processes between different oxygen atoms along with the movement of the polymer segment, thereby realizing the migration of Li+ in the PEO matrix [10,11,12,13].

Because PEO is a semi-crystalline polymer, and it is generally believed that the conduction of Li+ mainly occurs in the amorphous region of PEO [6, 14, 15], nanoparticle fillers (Al2O3, SiO2, TiO2, etc.) are usually added to inhibit the formation of crystalline phases in PEO as a method to increase the ionic conductivity [16,17,18,19]. However, the promotion in ionic conductivity by these nanoparticle fillers has a certain limitation. For example, adding too much fillers will form a solvent effect which will adversely reduce the ionic conductivity of the composite electrolyte [18, 20]. Researchers also studied the effect of different morphologies of fillers on PEO solid-state electrolytes. It’s found that, compared with nanoparticle fillers, one-dimensional nanowire fillers can better reduce the crystallinity of PEO, and the nanowires are more likely to form long-range ion transport channels at their interfaces, which is more favorable for the ion migration [21,22,23]. For example, Sheng et al. [24] reported that the addition of magnesium borate nanowires can enhance the ionic conductivity as well as the electrochemical window and mechanical properties of PEO solid-state electrolytes. The interaction between magnesium borate and anion bis(trifluoromethane)sulforimide (TFSI−) increased the ionic conductivity to 3.7 × 10–4 S·cm−1 at 50 °C. Miller et al. used halloysite nanotube (HNT) as fillers to prepare the composite electrolyte [25]. The positively charged HNT on the surface can promote the dissociation of lithium salt, and the ionic conductivity reached 1.11 × 10–4 S·cm−1 at 25 °C after addition of 10% HNT. Lu et al. prepared a salt-rich composite electrolyte by introducing Li6.75La3Zr1.75Ta0.25O12 (LLZTO) nanofibers into a salt-rich PEO6-LiTFSI solid-state electrolyte, which showed an ionic conductivity of 2.13 × 10–5 S·cm−1 at 25 °C [22]. These work all indicate that the interface of nanowires is essentially important for the long-range ion conduction in PEO matrix.

In our previous work [26], we reported that the oxygen atoms in the hydroxyl group could be complexed with Li+ through the phenolic resin nano-spheres fillers, and interact with the PEO segment to achieve Li+ transport. In order to further improve the conductivity of PEO-based solid-state electrolyte, modification of the hydroxyl or/and other oxygen-containing groups on the surface of the fillers is still necessary. In this work, we use glass fiber (GF) as the nanowire filler, which is further modified by hydrothermal carbon spheres (HCS) on its surface. Glass fiber is an inexpensive and easily available long-range fibrous material, while the hydrothermal carbon spheres are rich in oxygen-containing groups on the surface [27], such as hydroxyl, ester and other oxygen-containing hydrophilic groups. The hydroxyl-rich HCS were grown on the surface of glass fibers and then were added into the PEO system as a functional filler (GF@HCS). The electrochemical properties of the composite electrolyte were studied in detail. It is believed that the hydroxyl groups on the surface of the modified GF@HCS fillers can build plenty of long-range Li+ transport channels, which greatly promoted the Li+ conduction in the PEO-based electrolyte.

2 Experimental

2.1 Preparation of GF@HCS

The preparation process of GF@HCS composite electrolyte is illustrated in Fig. 1. Typically, 0.42 g glass fiber filter membrane (Whatman) was added to 200 ml deionized water for ultrasonic dispersion, and then 1 ml 20 wt% poly(diallyldimethylammonium chloride) solution (PDDA, Aladdin) was added, stirred for 2 h, and then filtered to obtain the dispersed glass fibers. For the growth of HCS, the dispersed glass fibers were added to a polytetrafluoroetylene (PTFE) hydrothermal reactor followed by addition of 1 g glucose and 50 ml deionized water. The hydrothermal reactor was sealed and placed in an oven at 180 °C for 24 h. After cooling, it was washed thoroughly with deionized water and ethanol and then filtered. The obtained brown GF@HCS product was finally dried in vacuum at 100 °C for 24 h before use. Additional, pure HCS was synthesized under the same conditions without the addition of GF.

Fig. 1
figure 1

Schematic diagram of composite electrolyte preparation and mechanism of Li+ conduction

Full size image

2.2 Composite solid-state electrolyte

PEO (60w, BASF) and LiClO4 (Aladdin) with the molar ratio of 16 (denoted as PEO16) were dissolved in anhydrous acetonitrile (Aladdin) to form a solution (10 wt%). Then different amounts of unmodified GF or GF@HCS (2 wt%, 4 wt%, 6 wt%) were added into the above solution. After sufficient stirring, the solution was placed in the PTFE mold to volatilize the acetonitrile solvent. The as-formed film was then dried in vacuum at 50 °C for 24 h to remove the residual solvent, resulting in the final composite solid-state electrolyte membrane.

2.3 Assembly and test of solid-state batteries

Firstly, the composite binder was prepared. PEO and LiClO4 with the molar ratio of 16 were added into the N,N-dimethylformamide (DMF, Aladdin) solvent, then PVDF powder was added, and the weight ratio of PEO–PVDF was kept at 7:3. Then the mixture was stirred and dissolved to form a solution with a weight ratio of 10%. Then, according to the weight ratio of 6:2:2, LiFePO4, composite binder and acetylene black were added into DMF solvent, mixed and stirred, and then the slurry was coated on aluminum foil, vacuum dried at 60 °C for 24 h. The loading density of LiFePO4 was controlled at about 1.0 mg·cm−2. The LiFePO4 solid-state cells were assembled in 2032-type coin cell with lithium metal sheet as the negative electrode and using the as-prepared composite electrolyte membrane as the middle layer. The voltage range of for the charge/discharge test was 2.5–3.8 V, whereas the charging and discharging current was set as 0.5C, respectively (1.0C = 150 mAh·g−1). Rate characteristics of the cells were tested beginning with 0.5C cycling and then followed by cycling at 0.5C, 1.0C, 2.0C, 4.0C and finally returned to 0.5C, respectively.

2.4 Material characterization and electrochemical tests

The morphology of samples was observed by field electron-scanning electron microscopy (SEM, Zeiss GeminiSEM 500). The molecular structure information was detected by Fourier transform infrared spectroscopy (FTIR, Nicolet IS5). The thermal behaviors of composite electrolyte were evaluated by synchronous thermal analyzer (STA 449 F3, Netzsch). The melting temperatures (Tm) and apparent melting enthalpy (ΔHm) were determined from the endothermic peaks, and the degree of crystallinity (X) of composite electrolyte was calculated from the following equation:

$$ X = \Delta H_{{{\text{m}},{\text{PEO}}}} /\left( {f_{{{\text{PEO}}}} \times \Delta H_{{{\text{PEO}}}} } \right) $$
(1)

where fPEO is the weight ratio of PEO in the related composite electrolyte, ΔHm,PEO is the apparent melting enthalpy of the composite electrolyte and ΔHPEO is the heat of melting of 100% crystalline PEO, which is 213.7 J·g−1 [23, 24].

The composite solid-state electrolyte was punched into a circle with a diameter of 16 mm, and sandwiched between two stainless steel (SS) electrodes for the measurement of ionic conductivity, which was evaluated by an alternative current (AC) impedance spectrometer (Solartron 1260) in the range of 1 Hz–1 MHz over a temperature from 25 to 100 °C in 10 °C intervals on heating. Linear sweep voltammetry (LSV) test was conducted using an electrochemical workstation (CHI 660E) at a scan rate of 0.1 mV·s−1 to evaluate the electrochemical stability. The composite electrolyte was assembled in a lithium symmetric cell to evaluate Li+ transference number (\({{t}}_{{\text{Li}}^{+}}\)), which can be calculated by the following equation [28]:

$$ t_{{{\text{Li}}^{ + } }} = \frac{{I_{{\text{s}}} (\Delta V - I_{0} R_{0} )}}{{I_{0} (\Delta V - I_{{\text{s}}} R_{{\text{s}}} )}} $$
(2)

where I0 and Is are the initial and steady state current, respectively, ∆V is the applied constant potential difference (20 mV), and R0 and Rs are the interface impedance before and after polarization, respectively.

3 Results and discussion

3.1 Characterization of GF@HCS

The optical photos and SEM images of the dispersed GF and GF@HCS are shown in Fig. 2. From the optical photos, it is found that the color of the glass fiber changed from white to brown after the hydrothermal treatment, which is similar to the HCS reported previously [27, 29]. Moreover, from enlarged SEM images, it is observed that the surface of GF is very smooth, while the surface of GF@HCS is well-coated by a layer of ultrafine nanoparticles that are expected to be the HCS.

Fig. 2
figure 2

SEM images of a, b GF and c, d GF@HCS fillers (insets showing corresponding optical photos)

Full size image

It is known that glucose will be dehydrated and polymerized into carbon nanospheres under the hydrothermal condition. The surface of such HCS is enriched with oxygen-containing groups, such as hydroxyl, ester, ketone, as well as a small amount of alkene groups and phenyl groups. At present, it is generally believed that HCS is a core–shell configuration. The shell is mainly composed of oxygen-containing hydrophilic groups such as hydroxyl groups, and its carbon content is low; while the encapsulated core is mainly an amorphous carbon structure composed of oxygen-containing hydrophobic groups such as benzene ring, alkene and a small amount of ester groups [27]. To explore the nanoparticles coated on the surface of GF, FTIR curves of GF, GF@HCS as well as the pure HCS are compared (Fig. 3a). The broad infrared absorption peak at about 3430 cm−1 represents the stretching vibration of –OH. The two close peaks at 2920 and 2840 cm−1 are characteristics of the stretching vibration of CH2 (aliphatic hydrocarbon), and those at 1750 and 1550 cm−1 signify the stretching vibration of C=O and C=C, respectively [30,31,32]. From FTIR spectra, the peaks corresponding to the vibration of CH2 and C=O for pure HCS are clearly observed in GF@HCS, and the intensity of the –OH related peak in GF@HCS is much larger than that in GF, which can confirm successful modification of the glass fibers by HCS after the hydrothermal treatment. From the thermogravimetric curves in Fig. 3b, it can be seen that GF@HCS sample starts to decompose at about 300 °C, and the content of HCS grown on the GF is ~ 2.2 wt%.

Fig. 3
figure 3

a FTIR spectra of GF, GF@HCS and pure HCS; b thermogravimetric curves of GF and GF@HCS

Full size image

3.2 Characterization of composite electrolyte

GF and GF@HCS fillers were, respectively, added into PEO16 to form a solid-state composite electrolyte. SEM images of the solid-state electrolyte with 2 wt% fillers (PEO16 + 2%GF and PEO16 + 2%GF@HCS) are shown in Fig. 4. Both GF and GF@HCS fillers are distributed in the entire PEO matrix and no essential morphological difference are observed for the two composite electrolytes.

Fig. 4
figure 4

Surface and cross-sectional SEM images of a, c PEO16 + 2%GF and b, d PEO16 + 2%GF@HCS composite electrolyte; e, f optical photos of PEO16 + 2%GF@HCS

Full size image

The ionic conductivity of the composite electrolytes with different contents of fillers is shown in Fig. 5. It can be seen that over-addition of GF or GF@HCS fillers cannot continuously increase the conductivity of the electrolyte, which is similar to the reports on other nanoparticle fillers [1, 18]. The addition of 2 wt% is demonstrated as the best scenario, and the modification of GF seems to be beneficial for the ionic conductivity of the PEO-based electrolyte, especially at elevated temperatures. For example, at 60 °C, the conductivity of PEO16 + 2%GF@HCS is promoted to 4.4 × 10–4 S·cm−1, being 400% higher than that of PEO16 (8.9 × 10–5 S·cm−1), while for PEO16 + 2%GF, the conductivity at the same temperature is only 2.0 × 10–4 S·cm−1. However, the room-temperature ionic conductivity of PEO16 + 2%GF and PEO16 + 2%GF@HCS is very close, being around 2.5 × 10–7 S·cm−1. Eventually, the electrolyte with modified GF fillers (PEO16 + 2%GF@HCS) presents an ionic conductivity over twice that of the one with only GF (PEO16 + 2%GF) at 60 °C. These results reveal that the modification of GF with HCS can greatly enhance Li+ transport in the PEO-based composite electrolyte at elevated temperatures. Such enhancement should originate from the increased amount of hydroxyl groups on HCS.

Fig. 5
figure 5

Ionic conductivity of composite solid electrolytes with different amounts of a GF and b GF@HCS fillers

Full size image

In order to clarify the reason of the different ionic conductivity, differential scanning calorimetry (DSC) tests were conducted on different composite electrolytes and the results are shown in Fig. 6. By analysis of the DSC data, the physical properties of the composite polymers are calculated and listed in Table 1. We can see that after the addition of GF or GF@HCS fillers, Tm of all polymers is decreased, which is consistent with other literature reports [33]. However, X shows first an increase and then a decrease with the increase in the amount of GF and GF@HCS fillers, which may be caused by the aggregation of the micron-sized glass fibers, leading to an inhibition effect on the crystallinity of PEO [18]. As previously reported [1, 18], when a small amount of fillers are dispersed in the PEO matrix, the formation of spherulites in PEO can be effectively suppressed (meaning a decrease in crystallinity), while when the amount of fillers continues to increase, the fillers tend to aggregate and cannot be effectively dispersed in the PEO matrix, resulting in a weakening of the inhibitory effect of fillers, which is accompanied by an increase in crystallinity. The calculated X values of PEO16 + GF and PEO16 + GF@HCS are close to each other, which may be the reason that their room-temperature ionic conductivities are at the same level.

Fig. 6
figure 6

a, b DSC curves of different composite electrolytes

Full size image
Table 1 Physical parameters of PEO-based composite electrolytes calculated from DSC data
Full size table

Steady-state current method was used to evaluate the \({{t}}_{{\text{Li}}^{+}}\) of the PEO composite electrolyte at 60 °C. The electrochemical results are shown in Fig. 7a, b. The calculation reveals that \({{t}}_{{\text{Li}}^{+}}\) values of PEO16 + 2%GF and PEO16 + 2%GF@HCS electrolyte are 0.12 and 0.18, respectively, indicating that the surface modification of GF by HCS can effectively promote the migration of Li+. This conclusion is consistent with our previous work [26], and it is verified again that the oxygen atom in the hydroxyl group is helpful for the PEO matrix in terms of ion conduction. Figure 7c shows LSV curves of the PEO composite electrolytes with different fillers. It is observed that adding GF or GF@HCS fillers does not affect the electrochemical stability of the PEO16 polymer electrolyte. The decomposition voltage of both polymers is over 3.9 V, which enables their use in LiFePO4-based solid-state batteries.

Fig. 7
figure 7

Current relaxation curves of a PEO16 + 2%GF and b PEO16 + 2%GF@HCS composite electrolytes, and (insets) corresponding impedance diagrams before and after polarization; c LSV curves of composite electrolytes (all tests being performed at 60 °C)

Full size image

3.3 Cell performance

When lithium metal is used as the negative electrode, the energy density of the battery can be significantly increased. However, lithium metal has high reactivity and uneven deposition of lithium metal is easy to form lithium dendrites. The growth of lithium dendrites can cause short circuits and cause safety hazards. Therefore, it is also necessary to test whether the solid polymer electrolyte has good stability to lithium metal [34]. Figure 8 is potential-time curves of the Li/electrolyte/Li symmetric cells (0.1 mAh·cm−2) deposited under different current densities. It is found that the PEO16 electrolyte with no fillers has a short circuit after 300-h cycling at 0.05 mA·cm−2, while PEO16 + 2%GF and PEO16 + 2%GF@HCS electrolytes maintain good cycling stability. According to previous studies [24, 25], adding filler is able to improve the mechanical strength of the composite polymer electrolyte, which can be one reason of the improved cycling stability of these composite electrolytes. It is also seen that PEO + 2%GF@HCS has a relatively lower voltage polarization than PEO16 + 2%GF. For example, as the cycling lasts from 100 to 600 h, the polarization of PEO + 2%GF@HCS is kept as low as 0.074 V, while the polarization of PEO16 + 2%GF increases from 0.084 to 0.11 V. The less polarization of the symmetric cell is in consonance with the higher ionic conductivity of the PEO + 2%GF@HCS composite electrolyte at the elevated temperatures.

Fig. 8
figure 8

a, b Potential-time profiles of Li/electrolyte/Li symmetric cells with deposition capacity of 0.1 mAh·cm−2 at 60 °C

Full size image

The PEO composite polymers were also used as electrolyte in LiFePO4/Li cells to evaluate their rate performances, which are shown in Fig. 9. The discharge capacity of the cells is significantly increased after the addition of both fillers. The addition GF@HCS is found to result in superior performance of the solid-state battery compared to the addition of only GF (Fig. 9a). The capacity of LiFePO4/PEO16 + 2%GF@HCS/Li cell at 4.0C reaches 121 mAh·g−1, which is obviously higher than those of the LiFePO4/PEO16 + 2%GF/Li (112 mAh·g−1) and LiFePO4/PEO16 + 2%GF@HCS/Li (94 mAh·g−1) cells. The enhancement of the high-rate performance is attributed to the higher ionic conductivity of the GF@HCS composite electrolyte, which leads to the lowest concentration polarization of the solid-state batteries (Fig. 9b, c).

Fig. 9
figure 9

a Rate performances of solid-state LiFePO4/electrolyte/Li half-cells and b–d corresponding charge/discharge curves at different rates (all tests being performed at 60 °C)

Full size image

The cycling performances of the batteries were also evaluated, as shown in Fig. 10. The solid-state batteries using all three composite electrolytes show a certain attenuation in capacity during cycling, which is probably due to unoptimized assembly of the solid-state battery. In the all-solid-state batteries, the traditional PVDF (as the binder) is absent because it cannot conduct Li+. Instead, the ion-conductive PEO polymer is used as the binder in the electrode, but the adhesion of PEO polymer is poor, which may cause poor electrode/electrolyte contact and lead to fast degradation of the battery [35, 36]. Nevertheless, by comparing the cycling performances of the LiFePO4/PEO16 + 2%GF @HCS/Li and LiFePO4/PEO16 + 2%GF/Li cells, it can still be verified that the modification of GF by HCS is a beneficial strategy to improve the electrochemical performance of the PEO-based solid-state battery.

Fig. 10
figure 10

Cycling performances of of LiFePO4/electrolyte/Li all-solid-state batteries at 0.5C and at 60 °C

Full size image

4 Conclusion

In this work, the positive effect of hydroxyl group on the conduction of Li+ is evidenced in PEO-based solid-state electrolyte using HCS modified GF as a functional filler. With optimum addition of the GF@HCS filler (~ 2 wt%), the ionic conductivity of the PEO composite electrolyte is increased from 8.9 × 10–5 to 4.4 × 10–4 S·cm−1 at 60 °C, being over twice that of the PEO electrolyte with only GF as the filler. Electrochemical tests reveal that the PEO16 + 2%GF@HCS composite electrolyte exhibits a high stability to lithium metal interface, and the LiFePO4-based solid-state battery shows excellent rate performance at elevated temperature. This work opens a new way to modify one-dimensional nanowire fillers for the PEO-based solid-state electrolyte.