Solidifying High-Concentration Electrolytes Using Faujasite as Nanosized Porous Zeolite Additive for Solid-Type Batteries

Abstract

A hybrid electrolyte composed of a high-concentration electrolyte made of dimethyl carbonate and lithium bis(fluorosulfonyl)imide and lithium exchanged faujasite-type zeolite exhibits solidification of the electrolyte with high ionic conductivity and excellent contact properties with a lithium metal anode and an olivine-type LiFePO₄ cathode. The assembled battery showed excellent cyclability for at least 100 cycles at a charge rate of 1 C at 60 °C with a discharge capacity retention of 98.6 % and high coulombic efficiency.

1. Introduction

In the face of the ongoing climate crisis, lithium-ion batteries play a pivotal role in the process of electrifying the energy grid in a move to eliminate the dependence of on fossil fuels, by serving as energy storage system for renewable energy generated by wind, solar or water. Traditional liquid electrolytes such as ethylene carbonate (EC), dimethyl carbonate (DMC) etc., with lithium-salt concentrations in the molar concentration range of 1 M (mol dm⁻³) are the standard in today’s lithium-ion batteries, exhibit high ionic conductivities in the 10⁻²–10⁻³ S cm⁻¹ order.¹ However, they are susceptible to leakage due to their high volatility, which, in combination with the high flammability of organic electrolyte solutions, leads to safety concerns. Combined with the incompatibility of working with Li-metal anodes due to the highly reductive potential, new electrolyte systems need to be developed.^2–4 In recent years, the use of high-concentration liquid electrolytes (HCLE) shifted into the focus of battery researchers due to their stability against Li-metal anodes, their high transference number of Li⁺ and their wide electrochemical stability window.^5–7 However, traditionally used separators made from polypropylene (PP) suffer from low wettability with high-concentration electrolytes.⁸ Further, their organic composition makes them susceptible during thermal runaways of batteries which lead to fatal malfunctions.⁹ Therefore, there is an interest in finding new separator materials with excellent wettability that can utilize the unique properties of high-concentration electrolytes while simultaneously increasing its performance by enhancing its safety properties. One approach is to utilize safe and wettable inorganic materials in separators, either as a blend¹⁰ or coating¹¹ on organic polymers or fully inorganic¹² separators. This strategy reduces the overall content of organic materials in the separator and increases the safety performance of the cell.

A promising group of inorganic materials are zeolites, 3-dimensional frameworks made from aluminosilicates with protons or metal-ions associated to it for charge balance.¹³ The aluminosilicates show, depending on the zeolite type, a variety of pore sizes and channel structures through which size selective mass transport can be facilitated. The combination of high surface area accompanied with Lewis-acidic properties makes zeolite a popular and widely used material in energy related fields.¹⁴

One possibility to combine the advantage of both are quasi-solid-state electrolytes (QSSE), consisting of a solid matrix combined with a liquid electrolyte. The role of the solid-matrix is to add stability, prevent leakages and enhance the safety of the system, while the liquid component facilitates the ion-transport.^15,16 Some groups reported the combination of lithium-concentrated liquid/polymer electrolyte with silicates/zeolites,^17–19 however, these systems suffered from the chemical/mechanical compatibility between electrolyte and electrodes, flexibility, or the rate capability while succeeded increasing the transference number of lithium ion and indicating the lithium-ion conduction through the zeolite pore by density functional theory (DFT) calculation.

In this report, we propose the use of nanosized lithiated zeolite with faujasite-group structure (FAUX-Li) powders as solidification agent of HCLE made from DMC and LiFSI for the preparation of a QSSE in lithium metal batteries. The combination allows stable reversible operation of cathode and anode. Furthermore, the unique structure of the highly Lewis acidic faujasite made of sodalite cages forming a supercage with high Si to Al ratio promises strong interaction between the HCLE and FAUX-Li, and simultaneously higher transference number of lithium ion. Combined with the easy preparation method of the electrolyte through simple mixing, the future potential when considering large scale manufacturability for industrial applications is highlighted.

2. Experimental

2.1 Preparation of electrolytes

The HCLE was prepared by mixing lithium bis(fluorosulfonyl)imide (LiFSI, IONEL^®, Nippon Shokubai Ltd.) with dimethyl carbonate (DMC, >99.5 %, Kishida Chemicals Ltd.) in a molar ratio of LiFSI : DMC = 1/1.5. After mixing the solution was left over night to allow for the complete dissolution of LiFSI. For the preparation of the quasi-solid-state electrolyte (QSSE), lithiated faujasite (FAUX-Li, Nakamura-choko Ltd., Si/Al = 1.15, average particle size of 100 nm) was mixed into the HCLE in a weight ratio of 1 : 3 and mixed hereafter via a shaker until the mixture showed a smooth texture. The hybrid electrolyte was left resting over the night before usage to confirm that no phase separation occurred.

2.2 Preparation of electrodes

For the preparation of the symmetric Li||Li coin cell, lithium metal sheets were stamped into circles with a diameter of 1.6 mm and pressed onto the coin cell. The hybrid electrolyte was filled into the coin cell and hereafter sealed to be used for further electrochemical measurements. For the preparation of the half-cell LFP cathode, lithium iron phosphate (LFP, MSE Supplies LLC) and acetylene black (AB, HS-100, Denka Ltd.) were mixed in a mortar for 10 min to ensure a homogeneous mixture. Hereafter, polytetrafluorethylene (PTFE, 6J, Chemours-Mitsui Fluoroproducts Ltd.) was added and kneaded with the pistol in the mortar until a smooth plastic like substance was formed. The dry cathodes were prepared with an LFP : AB : PTFE weight ratio of 70 : 25 : 5. Circular shaped cathodes were stamped out and pressed onto an Al-mesh welded onto an Al-sheet to avoid the possibility of the cathode moving during coin-cell assembly.

2.3 Electrochemical measurement

Lithium ion transference number (LiTN) were determined via the Evans-method in a Li||Li at 25 °C with impedance spectra taken in a frequency range of 0.1 MHz to 0.1 Hz (VSP-3e, Biologic) with an applied potential of 10 mV.²⁰ Conductivity of the electrolytes were determined with impedance spectra taken in a frequency range from 0.1 MHz to 0.1 Hz in a temperature range from 25 to 60 °C in a SUS||SUS configuration. Cyclic voltammetry (CV) and linear sweep voltammetry (LSV) were performed with a sweep rate of 0.1 mV s⁻¹ in Cu||Li (OCV to −0.5 V) and Li||Al (OCV to 6 V), respectively. Plating and stripping experiments (HJ1001SD8, HOKUTO DENKO) were conducted with Li||Li coin cells in a temperature regulated oven at 25 °C.

All the Li||LFP batteries were assembled using SUS clad coin cell cans (Hosen Corp., 2032). The detailed structure of the coin cell configurations for the liquid electrolyte and quasi-solid state electrolyte used for the conductivity measurement, LiTN measurement and half-cell measurement are displayed in the supporting information in Figs. S1 and S2.

2.4 Characterization and analysis

Scanning electron microscopy (SEM, JOEL JCM 7000, 15 kV) for the morphology analysis of the lithium metal anode was conducted before/after plating and stripping experiment. Thermogravimetric and differential thermal analysis (TG-DTA, Rigaku, TG-DTA8122 Themo plus EV02) between room temperature and 350 °C with temperature ramp of 5 °C min⁻¹ under Ar gas flow (20 ml min⁻¹) were conducted for investigating the thermal stability of the liquid/quasi-solid electrolytes. Fourier-transform infrared spectroscopy (FT-IR, Bruker Alpha II ECO-ATR) was performed in an argon filled glovebox for the characterization of the electrolyte solvation structure.

3. Results and Discussion

3.1 Characterization of electrolyte

In order to characterize the electrolyte, basic properties such as the conductivity, electrochemical stability window as well as LiTN were measured. Figure 1a shows the conductivity of the HCLE and QSSE in dependence on the temperature. With addition of zeolite to the HCLE, the conductivity of the QSSE dropped from 1.97 to 0.62 mS cm⁻¹. This decrease in conductivity can be explained due to the interaction between the nanosized zeolite with the HCLE led to a solidification of the electrolyte increasing its viscosity as the different viscous texture can be seen in Fig. 1c. Figure 1b shows the LiTN of the HCLE and QSSE. Current decay and EIS data are shown in Fig. S3. With addition of FAUX-Li, an LiTN increase from 0.50 to 0.67 for the QSSE was observed. The increase of LiTN can be explained by the interaction between the Lewis-acid sites of the FAUX-Li with the FSI⁻ anions of the Li-salt (structural detail of interaction discussed in Fig. 2), thus restricting its mobility in the QSSE, which was reported previously.¹⁷ Therefore, utilization of a FAUX-Li zeolite with a low Si/Al ratio is of advantage since it has been shown that lowering the Si/Al ratio increases the Lewis-acidity.²¹ By multiplying the LiTN with the conductivity of the electrolyte, the Li⁺ conductivity can be calculated, here described as σ_Li, leading to a conductivity of 0.98 and 0.42 mS cm⁻¹ for the HCLE and QSSE, respectively.

Figure 1.

(a) Arrhenius plots, (b) LiTN values, and (c) photo images of HCLE, FAUX-Li and QSSE, respectively.

Figure 2.

(a) Normalized FTIR signal of pure DMC solvent, HCLE and QSSE, (b) TG-DTA of HCLE and QSSE from 25 to 350 °C with a temperature increase rate of 5 °C min⁻¹.

FT-IR spectra of the QSSE and HCLE were compared to the pure solvent DMC in Fig. 2a. Free C=O vibration of pure solvent DMC in the purple line is characterized by the single emission band peak at 1753 cm⁻¹. When high LiFSI salt amounts were introduced such as the case in the HCLE (dashed black line), the free C=O emission band was drastically reduced and red-shifted to the shoulder peak at 1747 cm⁻¹, and a new emission band emerged at 1722 cm⁻¹ which can be assigned to the solvated C=O coordinating to the Li⁺. In the case of the QSSE, a slight reduction of the peak at 1722 cm⁻¹ and an increase of the peak at 1747 cm⁻¹, an indication for solvated C=O and an increase in free C=O was observed.²² This occurrence might be explainable by small amounts of dissociation between Li⁺ and DMC, therefore, increasing the free DMC amount, and simultaneous ion exchange of dissociated Li⁺ to the Lewis acidic sites in FAUX-Li. This explanation is supported by the possible lithium-ion conduction in the faujasite zeolite calculated by DFT calculations.¹⁸ On the other hand, removal of the DMC coordinated lithium ion by adding zeolite would indeed increase the free C=O signal. The slight increase in free DMC could further explain the slightly lowered electrochemical stability (discussed in section 3.2.). However, further investigations will be needed to clarify the liquid electrolyte-zeolite interactive behavior in a more systematic manner.

TG-DTA (Fig. 2b) was performed to investigate thermochemical properties of the QSSE and compared to the HCLE. The HCLE shows a continuous decrease in mass which can be attributed to the progressing evaporation of DMC with an increasing temperature. In comparison, the QSSE shows a lower mass loss even at high temperatures. When reaching 160 °C a weight loss of 30 % was observed and did not decrease noticeable afterwards. An explanation for the increased thermal stability might lie in the interaction of the zeolite with DMC through pore infiltration which would increase the solvent retention.

3.2 Electrochemical characterization

Plating and stripping behavior of the HCLE and QSSE were investigated in Cu||Li cells in a CV test and are shown in Fig. 3a. Both electrolyte systems show a stable plating and stripping of the lithium ions on the lithium metal without any signs of instability in the current profile. The lower current during the plating and stripping with the QSSE when compared to the HCLE is caused by the high viscosity limiting the mobility of the Li⁺. Further, in order to assess until which anodic voltage the electrolyte is able to operate, LSV was employed (Fig. 3b). The HCLE shows oxidative stability until 5.5 V while the QSSE shows oxidative stability until 5.25 V, which indicate QSSE has less oxidative stability than HCLE. The difference of oxidative stability might come from the difference of oxidatively decomposable free DMC amount predicted from FT-IR spectra mentioned in section 3.1.

Figure 3.

(a) CV of a Cu||Li cell with QSSE or HCLE in the range of 2.5 V to −0.5 V with a scan rate of 0.1 mV s⁻¹. (b) LSV of a Li||Al coin cell with QSSE or HCLE for the measurement of the electrochemical stability window from OCV to 6 V at a scan rate of 1 mV s⁻¹.

In order to examine stability of the QSSE with lithium metal comparing HCLE, symmetric Li||Li were tested as shown in Fig. 4. As can be seen the Li||Li coin cell with the QSSE is stable for 500 cycles with only a slow increase in voltage over time. On the other hand, the HCLE, although showing a lower initial voltage when comparing to the QSSE, showed an instable voltage profile above 125 cycles. This shows the good stability of QSSE with the Li metal. Figures 4b–4c shows the Li-metal surface of the symmetric coin cell before and after cycling. As shown in the images the surface after cycling showed smooth ball-like morphology of the lithium without any significant development of dendrites or sharp irregularities. Other published reports showed that high concentration electrolytes will lead to a smooth surface morphology of the lithium surface.²³ In strong contrast, the Li-metal surface of the Li||Li cell employing the HCLE, Fig. 4d, shows a strongly uneven deposition of the Li metal as well as cavities and pit holes, explaining the reason for the sudden rapid degradation after less than 125 cycles. One reason for the improved stability of the QSSE can be pointed to the increased lithium transference number, which allows for a more controlled stripping and platting of the Li during repeated cycling avoiding the creation of an uneven Li metal surface and therefore dendrite formation as was shown by previous reports.^24,25 Further, reports have shown that an Al/Si rich SEI has positive effects on the stability of Li-metal anodes. Therefore, it seems plausible that partial decomposition of the zeolite on the surface could lead to the formation of more robust SEI.^26,27 However, more research will be needed in the future to understand the interaction between the zeolite and Li-metal surface during SEI formation.

Figure 4.

(a) Galvanostatic cycling of symmetric Li||Li cells made from HCLE and QSSE with a charge/discharge time of 1 h at 0.1 mA cm⁻² at room temperature. (b) SEM image of the pristine lithium metal surface, (c) after cycling at 0.1 mA cm⁻² at 25 °C with QSSE, (d) SEM image of the lithium metal surface before and after cycling at 0.1 mA cm⁻² at 25 °C with HCLE.

3.3 Evaluation of half-cell performance

Performance of the QSSE was investigated with via galvanostatic cycling in a half-cell configuration consisting of LFP/Li-metal and compared to the HCLE. The charge-discharge curves of LFP||Li cells with QSSE before and after 100 cycles at 1 C-rate at 60 °C (Fig. 5a) displays the characteristic wide voltage plateau of LFP. After 100 cycles only a marginal capacity loss was visible, indicating the high stability of the QSSE in this system. Figure 5b shows the rate test of the QSSE ranging from 0.1 C to 1 C-rate, with negligible capacity loss when increasing the charge rate from 0.1 C-rate up to 0.5 C-rate.

Figure 5.

(a) Charge discharge curve of galvanostatic cycled LFP/Li half-cells with QSSE at a charge rate of 1 C at 60 °C before and after 100 cycles. (b) Charge rate test of LFP/Li half-cells with QSSE. LFP cathode with an active material loading of 1.5 mAh cm⁻². (c) Specific capacity, coulombic efficiency (CE) in dependence of the cycle number for galvanostatic cycled LFP||Li half-cells at 1 C-rate at 60 °C for HCLE and QSSE. LFP cathode with an areal capacity of 1.5 mAh cm⁻².

To compare the performance of the QSSE with the HCLE, galvanostatic cycling for 100 cycles at 1 C-rate at 60 °C was performed (Fig. 5c). While HCLE showed a high initial discharge capacity with 162 mAh g⁻¹, a continues capacity loss with increasingly instable coulombic efficiency was observed. After 50 cycles with a capacity retention of 139 mAh g⁻¹, the cell was unable to cycle further. The cycle curves of the HCLE, shown in Fig. S4, show unstable charging with overcharging beyond the theoretical capacity, indicating unstable Li-plating on the Li-metal anode. The visible charging spikes lead to the conclusion that internal micro-short circuits, caused by Li-dendrites penetrating the separator, are very likely the leading cause of the premature failure of the cell. This is further supported by the previously discussed low cycle life of the HCLE in the Li||Li experiment. On the other hand, the QSSE showed a decreased initial capacity of 151 mAh g⁻¹ when compared to the HCLE. However, it experienced negligible capacity loss after 100 cycles and excellent CE was maintained. This shows that the zeolite QSSE has good stability and excellent properties when used in a lithium-metal battery.

While the QSSE shows promising performance the conduction pathway of the lithium ions remains unclear. In future studies we are planning to conduct 2D-EXSY NMR and Li^6/7 isotope exchange NMR studies to investigate and clarify the Li-ion conduction pathway inside the QSSE.

4. Conclusion

In this work we demonstrated the effectiveness of the use of nanosized zeolites in combination with a high concentration carbonyl electrolyte for high performance lithium-metal batteries. We showed that the addition of lithiated FAUX-Li to the high concentration electrolyte, made from LiFSI and the carbonyl based solvent DMC, increased the stability with Li-metal anodes as shown by the stable cycling in a Li||Li coin cell for 500 cycles at 0.1 mA cm⁻². While HCLE showed rapid capacity degradation in LFP||Li half-cells, the use of the QSSE led to drastic improvement in cycle performance with a stable discharge capacity retention of 98.6 % (149 mAh g⁻¹) after 100 cycles.

Acknowledgment

We would like to thank Nippon Shokubai Ltd. for supplying the IONEL^® used for all experiments. We would also like to thank Hisao Ito for the experimental support. This work was supported by JST, the establishment of university fellowships towards the creation of science technology innovation, Grant Number JPMJFS2123.

Data Availability Statement

The data that support the findings of this study are openly available under the terms of the designated Creative Commons License in J-STAGE Data at https://doi.org/10.50892/data.electrochemistry.26493172.

• 1) A hybrid electrolyte composed of a high-concentration electrolyte made of dimethyl carbonate and lithium bis(fluorosulfonyl)imide and lithium exchanged faujasite-type zeolite exhibits solidification of the electrolyte with high ionic conductivity and excellent contact properties with a lithium metal anode and an olivine-type LiFePO₄ cathode. The assembled battery showed excellent cyclability for at least 100 cycles at a charge rate of 1 C at 60 °C with a discharge capacity retention of 98.6 % and high coulombic efficiency.

CRediT Authorship Contribution Statement

Schanth Hacatrjan: Data curation (Lead), Investigation (Lead), Methodology (Lead), Writing – original draft (Lead)

Kosuke Nakamoto: Conceptualization (Lead), Supervision (Lead), Writing – review & editing (Supporting)

Izumi Yamada: Formal analysis (Supporting), Methodology (Supporting)

Naoki Inui: Supervision (Supporting)

Kazuhiro Shibahara: Methodology (Supporting), Supervision (Supporting)

Daisuke Inokuchi: Investigation (Supporting)

Kohei Miyazaki: Supervision (Supporting), Writing – review & editing (Supporting)

Takeshi Abe: Funding acquisition (Lead), Supervision (Supporting)

Conflict of Interest

The authors declare no conflict of interest in the manuscript.

Funding

JST: JPMJFS2123

Footnotes

S. Hacatrjan: ECSJ Student Member

K. Nakamoto, I. Yamada, N. Inui, K. Shibahara, D. Inokuchi, and K. Miyazaki: ECSJ Active Members

T. Abe: ECSJ Fellow

References

• 1)  M. Dahbi, F. Ghamouss, F. Tran-Van, D. Lemordant, and M. Anouti, J. Power Sources, 196, 9743 (2011).
• 2)  C. Cao, Y. Zhong, and Z. Shao, Chin. J. Chem., 41, 1119 (2023).
• 3)  X. Wu, K. Song, X. Zhang, N. Hu, L. Li, W. Li, L. Zhang, and H. Zhang, Front. Energy Res., 7, 65 (2019).
• 4)  X. Yu, R. Chen, L. Gan, H. Li, and L. Chen, Engineering, 21, 9 (2023).
• 5)  G. A. Giffin, Nat. Commun., 13, 5250 (2022).
• 6)  J. Wang, Y. Yamada, K. Sodeyama, C. H. Chiang, Y. Tateyama, and A. Yamada, Nat. Commun., 7, 12032 (2016).
• 7)  Y. Yamada, Bull. Chem. Soc. Jpn., 93, 109 (2020).
• 8)  L. Zhao, Y. Li, M. Yu, Y. Peng, and F. Ran, Adv. Sci., 10, 2300283 (2023).
• 9)  W. Lin, F. Wang, H. Wang, H. Li, Y. Fan, D. Chan, S. Chen, Y. Tang, and Y. Zhang, ChemSusChem, 15, e202201464 (2022).
• 10)  H. Liu, Z. Dai, J. Xu, B. Guo, and X. He, J. Energy Chem., 23, 582 (2014).
• 11)  J.-J. Woo, Z. Zhang, N. L. Dietz Rago, W. Lu, and K. Amine, J. Mater. Chem. A, 1, 8538 (2013).
• 12)  M. He, X. Zhang, K. Jiang, J. Wang, and Y. Wang, ACS Appl. Mater. Interfaces, 7, 738 (2015).
• 13)  N. Salahudeen, Chem. Afr., 5, 1889 (2022).
• 14)  E. Pérez-Botella, S. Valencia, and F. Rey, Chem. Rev., 122, 17647 (2022).
• 15)  Z. Chang, H. Yang, X. Zhu, P. He, and H. Zhou, Nat. Commun., 13, 1510 (2022).
• 16)  H. Huo and J. Janek, Natl. Sci. Rev., 10, nwad098 (2023).
• 17)  K. Rafiz and J. Y. S. Lin, Sustain. Energy Fuels, 4, 5783 (2020).
• 18)  Z. Ding, Q. Tang, Y. Liu, P. Yao, C. Liu, X. Liu, J. Wu, and M. Lavorgna, Chem. Eng. J., 433, 133522 (2022).
• 19)  H. Jamal, F. Khan, S. Hyun, S. W. Min, and J. H. Kim, J. Mater. Chem. A, 9, 4126 (2021).
• 20)  J. Evans, C. A. Vincent, and P. G. Bruce, Polymer, 28, 2324 (1987).
• 21)  T. K. Phung and G. Busca, Appl. Catal., A, 504, 151 (2015).
• 22)  Y. Wu, A. Wang, Q. Hu, H. Liang, H. Xu, L. Wang, and X. He, ACS Cent. Sci., 8, 1290 (2022).
• 23)  J. Qian, W. A. Henderson, W. Xu, P. Bhattacharya, M. Engelhard, O. Borodin, and J.-G. Zhang, Nat. Commun., 6, 6362 (2015).
• 24)  J. L. Schaefer, D. A. Yanga, and L. A. Archer, Chem. Mater., 25, 834 (2013).
• 25)  Y. Lu, M. Tikekar, R. Mohanty, K. Hendrickson, L. Ma, and L. A. Archer, Adv. Energy Mater., 5, 1402073 (2015).
• 26)  Z. Ju, C. Jin, H. Yuan, T. Yang, O. Sheng, T. Liu, Y. Liu, Y. Wang, F. Ma, W. Zhang, J. Nai, and X. Tao, Chem. Eng. J., 408, 128016 (2021).
• 27)  J. Y. Kim, A.-Y. Kim, G. Liu, J.-Y. Woo, H. Kim, and J. K. Lee, ACS Appl. Mater. Interfaces, 10, 8692 (2018).