A safe and superior propylene carbonate-based electrolyte with high-concentration Li salt

Introduction

Shortage of fossil energy and increase of energy consumption lead to an ever-increasing demands for electric energy storage and conversion systems. Especially, the fast development of electric vehicles and hybrid electric vehicles poses higher requirements for the power, energy, and safety performance of power sources. Among the various electric storage devices, lithium ion batteries may be the best choice due to their high-energy, high-voltage, long lifespan, and environmentally friendliness. Nonetheless, as compared with lead-acid and nickel metal hydride batteries that are based on aqueous electrolytes, lithium ion batteries have potential safety hazards because the adopted organic electrolytes always use linear carbonates like diethyl carbonate (DEC), dimethyl carbonate (DMC), and methyl ethyl carbonate (EMC) as dominating solvents, which are highly flammable. Therefore, exploration of novel high-safety electrolytes is one of the most critical issues that need to be addressed for a wider application of Li ion batteries. Under these circumstances, nonflammable or flame-resisting solvents [1–6] have been investigated as the main solvent or co-solvent of electrolytes to improve the safety performance. For instance, aqueous rechargeable lithium batteries [7–9] were developed recently due to their good safety. Ionic liquids (ILs)-based electrolytes [4, 10–12] were reported to possess good fire-retardancy properties attributed to the non-volatility and non-flammability advantages of ILs. However, the costliness and incompatibility of ILs with graphite anode [13, 14] (ILs cannot form SEI film on graphite and the cations of ILs intercalate into graphite easily) restrict their practical application. In addition, cations of ILs can be regarded as an impurity since they may affect the migration of working Li⁺ ions. Also, other organic solvents such as γ-butyrolactone [1], cyclic phosphate [2, 3, 5, 6], and halogenated cyclic carbonates [2], vinyl-tris-(methoxydiethoxy)silane [15] were found to be able to improve the fire retardancy capability. However, the associated drawbacks like high sensitivity to water, poor stability, or high viscosity of these solvents hinder their large-scale application in Li ion batteries.

Propylene carbonate (PC), as an easily available solvent, has received great attention due to its low melting temperature (–48 °C) and high Li salt solubility. Most importantly, its high boiling and flash points can help to increase the safety properties. However, co-intercalation of PC-solvated Li⁺ results in serious exfoliation of graphite layers [16–19]. In general, two approaches have been adopted to prevent co-intercalation of PC. One is based on the formation of superior SEI films on the graphite surface through addition of additives into the electrolytes [15, 20–27]. The other is decreasing the solvation number of PC molecules per Li⁺ ion [28–30], which also make the intercalation and de-intercalation of Li⁺ possible. For the latter method, we previously found that addition of ionic liquid into PC-based electrolyte allowed formation of LiC₆ intercalation compounds, which is because the cations of ILs can compete with Li⁺ to be solvated by the limited amount of PC molecules [30]. Dimethylsulfoxide was also demonstrated to be able to solvate Li⁺ preferentially, which lead to reversible insertion of Li⁺ in PC-based electrolyte [31]. Furthermore, high concentration of Li salt (2.72 M) dissolved in pure PC solvent also renders effective insertion of Li⁺ due to the decreased solvation number of PC molecules per Li⁺ ion [28]. Nevertheless, to the best of our knowledge, the fire-retardancy capability of PC-based electrolyte with high-concentration Li salts has not been reported so far. In this work, we demonstrate that, adding adequate lithium hexafluorophosphate (LiPF₆) into a PC-based solvent (PC/EC with a volume ratio of 1/1) can not only improve the fire-retardant properties significantly, but also lead to superior cycling stability for Li⁺ storage as compared with the commercial electrolyte.

Experimental

LiPF₆, EC and DEC are provided by Shenzhen Capchem Technolgy Company. PC (99.98 %, water content below 20 ppm) was purchased from Lsomersyn Company. The electrolytes were prepared by dissolving different amounts of LiPF₆ in the solvents. Given the solubility of LiPF₆ in PC/EC co-solvent (a volume ratio of 1/1), the adopted concentration of Li salt ranges from 0.5 to 2.5 M. For the electrolyte with pure PC as solvent, 3 M of LiPF₆ was used due to the higher solubility of pure PC. 1 M LiPF₆ dissolved in EC/DEC, which is the common composition of commercial electrolytes, is employed as a reference. The electrolytes compositions are listed in Table 1. All the solutions were prepared and stored in an argon-filled glove box (VACOMNI-LAB), in which the water content was kept below 0.5 ppm.

The flammability of the electrolytes was tested by counting the time they cost to ignite and self-extinct as reported in other works [2, 3]. A glass fiber ball was soaked in the electrolyte and then a flame of gas burner was brought under the ball until the ball was lit. The time of ignition was recorded between the flame was brought and the ball was ignited. The time between the ball was ignited and automatically extinguished was defined as self-extinguishing time. The self-extinguishing time normalized against liquid mass was used to quantify the flammability of the electrolyte. Each electrolyte was tested five times. SEIKO TG/DTA7300 equipment was used to test the thermal stability of these electrolytes in the temperature range of 30 °C–300 °C.

The negative electrodes were prepared by spreading a slurry mixture of spherical graphite (BTR AGP-8), and carbon black (Super P), and PVDF binder onto copper foil. Positive electrodes were made of LiFePO₄, carbon black and PVDF coated on aluminum foil. All the electrodes were made with a standard procedure of mixing and coating, and the electrodes were dried under 120 °C in a vacuum oven for 16 h before the cell assembly. CR2032-type coin cells were assembled for galvanostatic cycling, and Celgard 3501 separator with a thickness of 15 μm was placed between the cathode and anode. Both half and full cells were assembled to test the electrochemical stability of the as-prepared electrolytes.

Charge/discharge characteristics and cycling ability of the cells were investigated with LAND CT-2001A tester. Prior to the performance test, including the rate capability test and cycle life test, the half cells based on graphite anode were cycled between 2 V and 0.01 V at a rate of 0.05 C for four formation cycles. For the full cells, the formation cycles were performed in a potential window of 2.2 V–4.1 V at the charge/discharge rate of 0.05 C. ZAHNER ENNIUM was used to test the electrochemical impedance spectra (EIS). The frequency range was set from 100 KHz to 10 mHz, and the amplitude was 10 mV.

Results and discussion

Table 1 shows the flammability results of the electrolytes. Because of the high volatility and inflammability of DEC, the electrolyte of 1 M LiPF₆/EC+DEC caught fire immediately when the gas burner was brought close and continued to burn for 124.1 s. Thus the commercial Li ion batteries with this electrolyte have latent safety dangers. When substituting the linear carbonate DEC with cyclic carbonate PC, the electrolytes took longer time to ignite and shorter time to self-extinguish, which can be attributed to the high boiling and flash points of PC. In addition, it can be concluded from Fig. 1 that the ignition time increases with the increase of Li salt concentration. When the salt concentration is higher than 1.5 M, the ignition time will be longer than 10 s, comparable to the flame-retarding ability of other electrolytes reported [2, 4]. Especially, the electrolyte 6 with 2.5 M of Li salt needs a long time of 26.8 s to ignite, and then extinguishes quickly. The ignition time of electrolyte 7 (3 M LiPF₆/PC) is slightly shorter than those of PC/EC co-solvent electrolyte in spite of its very high Li salt concentration, suggesting that the EC and PC cyclic carbonates may have a synergistic effect on fire-retardancy. As we expected, the self-extinguishing time of the electrolytes is inversely proportional to the ignition time. The above results reflect the excellent flame-retardant properties of PC-base electrolytes via use of high-concentration Li salt. The reasons can be ascribed to the decreased partial pressure of the flammable gas (the vapor of organic solvent) around the flame through addition of sufficient lithium salts.

Fig. 1

Comparative results of electrolyte flammability test.

Thermal stability of the electrolytes is another important parameter we should value because the batteries may be subjected to a high temperature environment when they are abused or charged/discharged at a large current. Figure 2 shows the thermogravimetric (TG) curves of the electrolytes used in this work. In the case of the reference electrolyte (1 M LiPF₆/EC+DEC), the mass begins to decrease at 50 °C, and a sharp mass loss observed between 50 °C and 130 °C should correspond to the evaporation of DEC. By contrast, PC-based electrolytes exhibit much better thermal stability. The electrolytes 2, 3, and 4 begin to lose mass at temperature of about 120 °C, while electrolytes 5, 6, and 7 begin to lose mass at about 130 °C. Evaporation of the electrolyte could lead to a large internal pressure in the batteries, which may make the batteries bulge or crack. From this point of view, PC-based electrolytes, especially those with high-concentration Li salt, are obviously more reliable than the commercial electrolytes.

Fig. 2

Thermogravimetric analysis of the electrolytes.

The initial galvanostatic discharge/charge curves of graphite anodes in the different types of electrolyte are shown in Fig. 3. It has been widely acknowledged that co-intercalation of PC with lithium ion will destruct the structure of graphite. In this work, similar to the work reported by Jeong et al. [28], we find that even for the pure PC solvent, intercalation/deintercalation of Li⁺ occurs reversibly when the concentration of LiPF₆ is up to 3 M. The reason should be due to the reduced salvation number of PC molecules per Li⁺ in the high-concentration electrolytes. The PC/EC co-solvent based electrolytes allow for reversible intercalation/deintercalation of Li⁺ under all the selected Li salt concentrations, which can be ascribed to the film-forming ability of EC on graphite surface as well as the reduced solvation number of PC molecules per Li⁺. For the electrolyte 2 with only 0.5 M of LiPF₆ added, the reversible capacity of graphite is considerably lower as compared to the commercial electrolyte 1 due to deficiency of lithium ion [32] and the irreversible capacity caused by the intercalation of PC. When the concentration of Li salt is higher than 1 M, the reversible capacities of graphite approximate to the cells based on electrolyte 1, suggesting that high-concentration of Li⁺ salt also enables effective insertion-extraction of Li⁺.

Fig. 3

The initial discharge/charge curves of Li//graphite cells with the seven types of electrolyte.

Electrochemical impedance spectroscopy was used to investigate the effect of high-concentration electrolyte on the impedance properties of the battery systems. As can be seen from Fig. 4, the Nyquist plots consist of one or two depressed semi-circles at mid-high frequency and an inclined line at low-frequency. The intercept of the point with real axis at high frequency reflects the resistance of bulk electrolyte (R_b). It can be roughly concluded that R_b increases gradually with the increase of Li salt concentration, which can be due to the high viscosity of the electrolytes at high Li salt concentration. The intercept of the depressed semicircles on real axis is representative of the resistance of charge transfer process (R_ct) and SEI film. Appearance of a single semicircle is most likely due to the proximity of the reaction time constants between the SEI film and charge transfer process [11]. For the electrolyte with a low concentration of 0.5 M, the semicircle is extremely large probably due to the shortage of Li⁺ and serious reduction reaction of solvent. As a result, the charge-transfer reaction is difficult to proceed and a quite thick SEI film may be formed. For the PC/EC based electrolytes with Li salt concentration higher than 1 M, diameters of the semicircles are slightly smaller as compared with the commercial electrolyte 1. In addition, it can be observed that diameters of the semicircles decrease slightly as the concentration increases. This is probably because of the promoted charge transfer of Li⁺ and formation of a thinner SEI film due to the presence of abundant Li⁺ and reduced solvent molecules. For the electrolyte 7 with 3 M of LiPF₆ added into pure PC, the semicircle is also extraordinarily large, which probably results from the poor wettability of the electrolyte due to its high viscosity.

Fig. 4

EIS spectra of Li//graphite cells with the seven types of electrolyte.

Rate behavior of the Li//graphite half cells is displayed in Fig. 5. Among the different types of electrolyte, the electrolytes 2 and 7 show the worst rate performance resulting from their extremely large resistance as indicated by the EIS results. For the PC/EC based electrolytes with high Li salt concentration, especially for those with 2 and 2.5 M of LiPF₆ added, their rate capability is slightly better than the commercial electrolyte 1. Since the resistance of bulk electrolyte is large at high concentration, their good rate performance should be mainly ascribed to the low charge transfer and SEI resistance.

Fig. 5

Rate performance of Li//graphite cells with the seven types of electrolyte.

Cycling performance of the Li//graphite half cells is shown in Fig. 6. For the PC based electrolyte with high concentration, their cycling stability is slightly better than the commercial electrolyte 1. On the contrary, the capacity declines drastically during the extended cycles for the electrolyte 2 with low concentration, which is believed to relate to the deficiency of Li⁺ [32]. Considering that deficiency of lithium ions is more serious in full cells as compared with those in half cells, we also investigate the cycling performance of graphite//LiFePO₄ full cells with these high-concentration electrolytes. As shown in Fig. 7, cells with 1.5 M and 2 M of LiPF6 in PC/EC display the best cycling stability. Remarkably, cycling performance of these two electrolytes is even superior to the commercial electrolyte 1, which can be ascribed to the presence of abundant Li⁺ that can shuttle back and forth between cathode and anode. For the electrolyte 6 with 2.5 M of LiPF₆, the cycling performance is slightly worse than those with 1.5 M and 2 M of Li salt, which may result from its high viscosity, poor wettability, or large electrolyte resistance. Based on the above results, we conclude that the electrolyte 5 with 2 M of LiPF₆ is the best choice for Li-ion batteries taking the safety and electrochemical performance into account.

Fig. 6

Cycling performance of Li//graphite cells with the seven types of electrolyte.

Fig. 7

Cycling performance of graphite//LiFePO₄ full cells with seven types of electrolyte.

Conclusion

In conclusion, we present a high-safety PC-based electrolyte via use of high-concentration Li salt. As compared with the commercial electrolyte with 1 M of LiPF₆ dissolved in EC/DEC, the PC/EC based electrolytes with high concentration of Li salt (2 M) not only possess superior fire-retardancy capability and thermal stability, but also manifest excellent cycling performance for applications in both Li//graphite half cell and graphite//LiFePO₄ full cell. Therefore, this new type of PC-based electrolyte shows great prospects for practical applications in Li-ion batteries.