Polyurethane/Li₁₀GeP₂S₁₂ composite electrolyte with high ions transfer number and ions capture for all-solid-state lithium batteries

INTRODUCTION

All-solid-state Li-ion batteries (ASSLBs) have been regarded as an alternative to traditional liquid Li-ion batteries (LLIBs) for their higher energy density, better safety, and flame retardancy^[1]. The core component of ASSLBs is the solid-state electrolyte (SSE), which is linked to the overall performance of the battery^[2]. Among all types of SSEs, Li₁₀GeP₂S₁₂(LGPS)^[3] exhibits a high ionic conductivity (10^-2 S/cm), which is comparable to liquid electrolytes. However, the main drawbacks of inorganic solid electrolytes are their poor flexibility and high interfacial impedance between electrolyte and electrodes^[4,5]. It has been discovered that the presence of high interfacial impedance reduces the capacity retention and operating efficiency at high C-rates and long cycles, lowering the battery's overall performance^[6,7].

Additionally, solid polymer electrolyte (SPE) has many advantages, such as flexibility and good film formation ability^[8], which can compensate for the disadvantage of inorganic solid electrolytes. Therefore, preparing polymer/ceramic composite electrolytes is an effective strategy for overcoming the drawbacks of inorganic ceramic electrolytes. At present, physical mixing with different polymer/inorganic ratios is a simple and easy method to screen the optimal composite electrolyte systems^[9-11]. Therefore, researchers paid little attention to the impact of the interaction of polymer and inorganic ceramic molecules on Li⁺ ions. In addition, as for SPE, the ionic conductivity, as well as the ionic transfer number, are also key factors for choosing the polymer substrate [most polymers have low ionic conductivity(10^-5~10^-6 S/cm)^[9] and the ionic transfer number(< 0.5)^[12].

In our previous work^[2], we prepared a PU-based SPE with excellent mechanical properties and high ionic conductivity at room temperature (2 × 10^-3 S/cm). Since the PU-based SPE has a unique soft and hard segment structure, Li⁺ ions primarily transport in polyether polyols (PPG), so the soft segment in the PU-based SPE has the ability to dissolve lithium salts^[13-15] and disperse ceramic particles^[16].

In this study, a composite electrolyte [(PU-LGPS)/Li⁺, PLL] was created by modifying the soft segment functional groups in a PU-based electrolyte and incorporating LGPS particles (1~5 µm) into the polymer systems. In this study, Fourier transform infrared (FTIR) spectroscopy, Raman spectra, and X-ray photoelectron spectroscopy (XPS) were used to characterize the structure and properties of PLL. Meanwhile, we used the electrochemical impedance spectroscopy (EIS) test and assembled the battery to evaluate the specific charge/discharge capacity and electric cycle stability at room temperature. Furthermore, using first-principles simulation, we investigated the mechanism of the composite system that improves Li⁺ transport.

RESULTS AND DISCUSSION

Figure 1 depicts the PLL preparation process. Supplementary Material Section 1 shows all materials and preparation details. PPG (20 g, 0.03 mol) first reacts with LiOH·H₂O (5 g, 0.1 mol) to change the functional groups from “-OH” to “-OLi”. The mixture was then supplemented with LiTFSI (1 g, 3 mmol) and Al₂O₃ (0.05 g, 0.4 mmol) and stirred for 2 h until fully dissolved. Following that, LGPS (5 g, 8 mmol) particles were added, and the solution was evenly stirred to form a homogeneous solution. Finally, the 7.5 mmol MDI was added to form PLL. We can disperse the LGPS primarily in soft segment chains using this method, and the interaction of Al₂O₃ with functional groups has been described in our previous work^[2].

Figure 1. The preparation process of PLL [(PU-LGPS)/Li⁺]

Infrared and Raman spectra were used to characterize the PU and the effects of different LGPS addition ratios on functional groups and chemical bonds. Infrared spectra characterized the chemical component and internal bonding behavior of PU, and the change in peak position after the addition of LGPS (a detailed analysis of the analyzed infrared spectra) is shown in Supplementary Material Figure 1A^[17-19]. Supplementary Figure 1B depicts the Raman spectra of PU and PU + LGPS. The Raman characteristic peaks at 275 cm^-1, 409 cm^-1, and 556 cm^-1[20-21] can be ascribed to the deformation vibration peak, stretching vibration peak, and asymmetric vibration absorption peak of PS₄^3- and P₂S₇^4-, respectively. The stretching vibration of LiS₄^3- causes the absorption peak at 341 cm^-1. Furthermore, the stretching vibration peak of the “P-S-P” bond can cause the absorption peak at 512 cm^-1[22]. The obtained results show that the composite material PLL has good chemical compatibility [Figure 2].

Figure 2. XPS spectra of electrolyte PLL. (A) full spectrum diagram; (B) C_1S, illustrations: O_1S; (C) Li_1S; (D) S_2P; (E) P_2P; (F) Ge_3d.

We used XPS characterization to further investigate the bonding behavior of the composite electrolyte, and the results are shown in Table 1.

Through the XPS analysis of the PLL, it can be known that the PLL’s internal state is stable.

In addition, the physical properties of PLL are also critical, which are associated with electrolyte thermal stability and mechanical properties. Furthermore, the surface morphology of PLL is shown to efficiently identify element distribution in PLL and PLL affinity for various fillers.

First, we examined the XRD patterns of PU and LGPS, respectively. The sample LGPS peak location [Supplementary Figure 2] is consistent with the simulation XRD patterns^[30]. The figure shows that the prepared PU material has a broad peak at 2θ near 19.5° and 46.0°, indicating that the prepared PU material has a certain degree of crystallinity^[31].

Figure 3A presents that the glass transition temperature (T_g) value was approximately -51 °C, indicating that the electrolyte has good flexibility at ambient temperatures. Figure 3B displays the stress-strain curve for PLL. It demonstrates that PLL has a stress strength of 4.5 MPa and an elongation-at-break level of 220.1%. These characterizations show that PLL has good stress strength characteristics, with higher stress-strain properties compared to other reported SPEs^[32]. Supplementary Figure 3 clearly shows an SEM image for surface morphology and element mapping for electrolyte PLL. Besides, the electrolyte surface is smooth and evenly distributed, and the elements found in reactant LGPS are present in this electrolyte system. Al₂O₃ nanoparticles and LiTFSI within the polymer system are evenly distributed in this system, indicating that these added fillers are compatible with the system^[2].

Figure 3. (A) DSC curve of the PLL; (B) the stress-strain curve of the as-prepared PLL.

Ionic conductivity is an important electrolyte parameter. Table 2 details the ionic conductivity changes caused by various (LGPS/PU = x)/Li⁺ ratios.

Although the ionic conductivity of PLL continues to rise with the increase of LGPS, for electrolytes, in addition to the ionic conductivity, the stress-strain performance is considered a vital index to evaluate the Li dendrite inhibition ability of the electrolyte.

Figure 4A shows the Arrhenius plots for the ionic conductivities of different proportions of (LGPS/PUx)/Li⁺. Through the slope of these curves, the relationship between the ionic conductivity and temperature of several groups of samples was established. The linear relation between temperature and conductivity demonstrates that PLL electrolytes obey the ion transport mechanism originating from the polymer chain vibration mechanism. Figure 4B explains the metric for screening the optimal electrolyte systems. Three samples with proportions of 30%, 40%, and 50% were selected. Since the conductivity of the sample with a 60% proportion is not much higher than that of the sample with a 50% proportion, and its mechanical property is also lower than that of the three samples (30%, 40%, and 50%), we did not consider the sample with 60%. The activation energies of the three different proportions of (LGPS/PU)/Li⁺ were 149.3 kJ/mol (30%), 64.6 kJ/mol (40%), and 34.7 kJ/mol (50%), respectively [Supplementary Figure 4]. The activation energy calculation details are shown in Supplementary Material Section 1. Figure 4B demonstrates the relationship between ionic conductivity and stress. The sample with an LGPS ratio of 40% has superior mechanical properties than that of the sample with 50%. As a result of extensive analysis, the optimal proportion of PLL is 40%.

Figure 4. (A) Arrhenius plots for the ionic conductivities of different proportions of (LGPS/PUx)/Li⁺; (B) Relationship between ionic conductivity and stress, PLL at different temperatures; (C) AC impedance; and (D) Arrhenius plots for the ionic conductivities (40%).

Figure 4C displays the AC impedance of a PLL at different temperatures. As the temperature rises from 25 °C to 80 °C, the conductivity rises from 3.1 × 10^-3 to 6.1 × 10^-3 S/cm. The ionic conductivity can be identified through polymer chain movement based on Li⁺ coordination and interaction based on the elevation of ionic conductivity as temperature increases^[33]. As shown in Figure 4D, the linearity of Arrhenius plots for ionic conductivity as a function of temperature suggests that polymer electrolyte conductivity follows Arrhenius law^[34].

This study determined the ionic transfer number of the composite to further demonstrate Li⁺ transfer kinetics. Figure 5 clearly shows the potentiostatic polarization curve and the variation of the impedance spectra before and after the polarization of the PLL composite Figure 5A. Furthermore, the value of T_Li is 0.56, and the ionic transference number calculation details are shown in Supplementary Material Section 2.

Figure 5. (A) Potentiostatic polarization curve of a symmetric lithium battery (Li|PLL|Li), the illustrations are the impedance comparison of a symmetric lithium battery(Li|PLL|Li) before and after polarization; (B) First-principles theory simulation analyzing the charge distribution in the LGPS crystal.

To investigate the charge distribution in the LGPS crystal, we used first-principles theory simulation^[35-37]. Figure 5B depicts the charge being primarily concentrated around “S” atoms, demonstrating the ease with which electron exchange can occur. Based on previous research, we demonstrated that “O” atoms can act as Li⁺ trapping sites due to dense negative charge distributions surrounding “O” atoms. These findings suggest that Li⁺ deintercalated from the LGPS lattice can be absorbed by “O” atoms on PU. Furthermore, the functional group modification process can reduce the coupling impact of “O” atoms on Li⁺, resulting in a moderate capture behavior. The absorption energy of the “-OLi” group towards Li⁺ is shown to be 0.55 eV in comparison to that of the “-OH” group (1.01 eV)^[2], indicating that the “-OLi” group can both capture and easily give out the Li⁺ from LGPS crystal. Therefore, the “-OLi” group can function as a Li⁺ transfer medium, and the “S” sites in the LGPS crystal, in conjunction with the “O” site on the PU molecule, can function as an interfacial ion transfer channel, contributing to the high ionic conductivity.

The LFP|PLL|Li battery was built to investigate the use of an electrolyte for battery device performance. Figure 6A depicts specific LFP|PLL|Li battery capacities at different cycle C-rates. At 0.2 C, the capacity was 167.7 mAh/g, which was very close to the theoretical specific capacity of LFP cathode materials (170 mAh/g). It can show that the PLL electrolyte is effective at transporting ions and contacting solid-state LFP|PLL|Li battery electrodes. Furthermore, our assembled battery has high stability in the charging/discharging platform under varying C-rates with no polarization, indicating that our prepared electrolyte significantly improves battery stability while improving electrolyte-electrode contact, reducing the phenomenon of battery polarization. Figure 6B depicts the battery charging/discharging process at various C-rates. When the cycling rate is changed from 0.2 C to 5 C and then back to 0.2 C, the battery capacity reaches 99%, indicating that the internal rate of the battery is stable. Figure 6C depicts a schematic plot of the battery during the bulb-lighting process.

Figure 6. (A) LFP|PLL|Li, the first charge/discharge curves for 0.2 C, 1 C, 2 C, 3 C,4 C, and 5 C at ambient temperatures; (B) Capacity with various C-rates; (C) Battery luminescence picture (LFP|PLL|graphite); (D) Long-cycling properties offered based on varying C-rates; (E) Voltage-time curves of the symmetric lithium battery (Li|PLL|Li) at the current density of 0.5 mA/cm² and 5 mA/cm², as well as 10 mA/cm².

Figure 6D depicts the long-cycle performance of the batteries based on varying C-rates. After 2,000 cycles, the battery retention rate can still be 95.7% at 0.2 C, 91% at 3 C, and 87.8% at 5C, indicating the assembled batteries' excellent performance stability.

The voltage profiles of the symmetric Li|PLL|Li cells at 0.5 mA/cm², 5 mA/cm^2, and 10 mA/cm² are shown in Figure 6E. The cells maintained stable voltage profiles for 1,250 h at 0.5 mA/cm². When the current is increased tenfold (5 mA/cm²), the voltage fluctuation increases from 0.01 V to 0.03 V. Despite the increased fluctuation range, the circulation remained stable within 1,000 h, indicating no obvious polarization phenomenon appeared. With the current density increased to 10 mA/cm², more than 700 h of stable circulation and hysteresis can be obtained, indicating the good stability of Li|PLL|Li batteries. Through the investigation of the electrolyte PLL and the overall performance analysis of the assembled batteries, which are superior to previous reports^[38-40] [Table 3]. We assembled an NCM|PLL|Li battery to verify that PLL has a universal adaptive electrode. The results [Supplementary Figure 5A] demonstrate that the PLL we developed can be adapted to the NCM electrode. Furthermore, 40 cycles were performed at 5 C to compare the two groups of batteries (NCM|PLL|Li battery and LFP|PLL|Li battery), and both groups of batteries had stable cycles, demonstrating the excellent performance of the PLL prepared by us [Supplementary Figure 5B].

As shown in Table 4, when the current density of the PLL is increased from 0.5 mA/cm² to 10 mA/cm², the voltage increases by 0.04 V. However, in some other relevant works^{[41,42,44,45]}, the current density was still less than 0.5 mA/cm². Despite the fact that Deng’s work^[43] increased the current density to 4.5 mA/cm², the voltage fluctuation value was 0.5 V. In contrast, there is no obvious polarization phenomenon in the PLL electrolyte. This result shows that the PLL electrolyte has a high intrinsic ionic conductivity. PU is an elastic electrolyte that can form a good contact interface with the electrode, promoting interfacial ion transport.

CONCLUSIONS

In conclusion, a new solid composite electrolyte PLL [(PU-LGPS)/Li⁺] was developed. LGPS particles are dispersed in soft segment structure (PPG) on PU, and functional groups “-OH” and ionic bonds “-OLi” interactions on PPG and LGPS are used to promote Li⁺ transport. The PLL has a high Li⁺ transfer number of 0.56 as well as increased ionic conductivity [3.1 × 10^-3 S/cm (25 °C), 6.1 × 10^-3 S/cm (80 °C). Moreover, we used the first-principles theory to confirm the enhancement mechanism of ion transport in the dual system. The batteries that we have assembled perform admirably. The specific discharge capacity at 0.2 C was nearly 167.7 mAh/g, which approaches the theoretical specific capacity of LFP materials (170 mAh/g), resulting in excellent capacity retention (95.7%) after 2,000 cycles at 0.2 C, as well as retention of 91% and 87.8% after 2,000 cycles at 3 C and 5 C, respectively. This research provides a solid theoretical foundation and experimental demonstration for preparing composite polymer electrolytes, which improve battery efficiency.