Development of Lithium Ion Conducting Liquid: Methylurea-Based Eutectic Electrolytes for Lithium Batteries

Abstract

Deep eutectic electrolytes (DEEs) are attracting increasing attention as liquid-state electrolytes for secondary batteries because they are potentially low cost, display low flammability, and are environmentally friendly. However, to date limited DEEs have been developed and explored for lithium-ion battery (LIB) applications, with most reports showing unsatisfactory capacity retention, a narrow potential window for battery operation, and an unstable solid electrolyte interphase (SEI) layer leading. Herein, we develop DEEs based on lithium bis(fluorosulfonyl)amide, LiFSA, and a series of urea derivatives as Li ion-conducting DEEs. Despite similar structures for the urea derivatives, i.e. methylated urea, we found that 1,3-dimethylurea (1,3-DMU) could form Li ion-concentrated DEEs across a wide range of LiFSA : 1,3-DMU ratio, while the LiFSA : urea DEE was liquid only in a limited range of molar ratios, i.e. LiFSA : urea close to 1 : 4 (mol/mol). By examining the electrolyte structure via Raman spectroscopy, we observed increased aggregation for DEE with higher LiFSA concentrations. We further confirmed non-flammability and electrochemical stability among the DEEs with potential windows ranging from ∼3.35 V for LiFSA : urea (1 : 4) to an impressive 6.42 V for LiFSA : 1,3-DMU (1 : 2) at a Pt foil electrode. During charge-discharge of Li₄Ti₅O₁₂ (LTO) electrodes, we observed good capacities and retention for the LiFSA : urea (1 : 4) and LiFSA : 1,3-DMU (1 : 2) DEEs. High Coulombic efficiencies (CEs) were achieved in the LiFSA : 1,3-DMU (1 : 2) DEE with its high LiFSA content that led to more substantial FSA-derived components in the SEI structures after cycling. We further tested positive electrode materials, including LiFePO₄ that showed excellent capacity retention and CEs near 100 % across 50 cycles. In all, we find that the dimethylurea-based DEEs show an opportunity for non-flammable and high-voltage Li batteries.

1. Introduction

Currently, lithium-ion batteries (LIBs) are in high demand as power sources for both small portable electronic devices, such as smartphones and computers, and the growing electric vehicle market. LIBs are ideal for portable energy storage because they show high energy densities, operate at high voltages (>3.0 V), and are lightweight on the basis of the light atomic weight and lower E⁰ of lithium.¹ Their operation tends to rely on organic electrolytes, e.g. lithium hexafluorophosphate (LiPF₆) in carbonate, an ester solvent, which show good electrochemical stability, high ionic conductivity, and a relatively wide potential window.² Such battery electrolytes enable high-voltage operation, but they are generally volatile, flammable, and toxic. These characteristics have remained a safety concern for commercial batteries and also a major hurdle for using Li metal as a high capacity negative electrode.³ Furthermore, thermal runaway and explosion risks are directly linked to reactions with the electrolyte.^4,5 In the interest of safety and expanding the performance of LIBs, researchers are looking to develop alternative electrolytes including nonflammable or flame-retardant organic solvents,^6,7 highly concentrated organic^8,9 and aqueous electrolytes,^10,11 and solid-state electrolytes.^12,13

Recently, deep eutectic electrolytes (DEEs) have been recognized as a new ion-conducting liquid as well as potential non- or low-flammable electrolytes that can be prepared with ease from low cost, safe, and environmentally benign components.^14–17 Generally, DEEs are a combination of hydrogen bond donor and acceptor compounds (or a Lewis base/acid combination) that are solid state at room temperature, but show a deeply depressed melting point upon mixing. For practical battery applications, e.g. vehicles and portable electronics, the melting point should be much lower than room temperature to account for changes in weather and environment. To date, room-temperature DEEs explored for LIBs include mixing lithium salts, e.g. lithium bis(trifluoromethanesulfonyl)amide (LiTFSA), LiPF₆, LiNO₃, lithium difluoro(oxalato)borate with N-methylacetamide (NMA),¹⁵ methylcarbamate,¹⁸ succinonitrile, or 1,3,5-trioxane.¹⁹ In addition, DEE has been extensively studied for application to Na-ion²⁰ and Al-ion²¹ batteries. However, some issues remain such as low discharge capacities and retention, poor Coulombic efficiencies (CEs), and low electrochemical stability of Li-insertion material. Also, a key feature of high-voltage LIBs, the solid electrolyte interphase (SEI) is not as stable as the SEI found in traditional organic electrolytes.¹⁸ This may be due to differences in the SEI composition and properties that are derived from the electrolyte components. We believe that many other DEEs still remain unexplored for application in LIBs despite a wide variety of potential candidates for improving SEI performance and cycling.

Of particular interest are DEEs based on urea, are being explored for application in a variety of fields.^22,23 Urea has desirable characteristics for use in DEEs and has already found a few applications for aluminum batteries.²⁴ Herein, we focus on developing DEEs by mixing urea, or methlated urea derivatives, with lithium bis(fluorosulfonyl)amide (LiFSA), for application in LIBs. LiFSA has a lower melting point (133 °C) than LiTFSA and LiPF₆ used in previous papers, and the FSA⁻ anion is widely known to induce a good passive layer on the negative elecrode that inhibits electrolyte degradation.^25,26 In this study, we found that despite the urea derivatives having similar carbonyl and amino functional groups, which are essential to depress the melting point,^15,18,27 not all of the urea derivatives were capable of forming DEEs. Further, the urea structure was shown to impact the electrochemical stability of the DEE and cyclability of both positive and negative electrodes. Lastly, surface analysis indicated differences in the developed SEI post cycling, which suggests opportunity for finding further improved DEEs for higher cyclability and higher voltage batteries.

2. Experimental

2.1 Electrolyte preparation

DEEs were prepared by mixing various molar ratios of LiFSA (≥99 %, Kishida Chemical) and urea derivatives at 25 °C ± 5 °C for ∼10 hours in an Ar-filled glove-box (dew point = −70 °C). Urea (≥99.0 %, Wako Pure Chemical Industries) and its derivatives, 1-methylurea (≥98 %, Combi-blocks), 1,1-dimethylurea (≥96 %, Kanto Chemical), 1,3-dimethylurea (≥98 %, Tokyo Chemical Industry), tetramethylurea (≥99 %, Sigma-Aldrich) were used as purchased without further purification.

2.2 Electrode preparation

The positive and negative electrodes consisted of 80 wt% active material, 10 wt% conductive carbon, and 10 wt% polyvinylidene fluoride (PVdF, #9100, Kureha). Acetylene black (AB, Li-400, Denka) was used as the conductive carbon for all battery electrodes. For the active materials, we used Li₄Ti₅O₁₂ (LTO) (Sigma-Aldrich) for the negative electrodes, and LiFePO₄ (LFP) (Sigma-Aldrich), LiMn₂O₄ (LMO) (TOSHIMA), LiNi_1/3Co_1/3Mn_1/3O₂ (NMC) (NIPPON CHEMICAL) for the positive electrodes. Both electrodes were prepared by casting the mixture N-methylpyrrolidone (NMP) slurry onto Ti foil with a doctor blade and drying at 80 °C under a vacuum. Activated carbon electrodes were prepared with 80 wt% activated carbon (YP50F, Kuraray), 10 wt% Ketjen Black (KB, Carbon ECP, Lion), and 10 wt% polytetrafluoroethylene (Daikin) on Ti mesh and dried at 100 °C. Additional LTO negative electrodes were prepared using 10 wt% carboxymethyl cellulose, CMC (CMC#2200, Daicel Miraizu Ltd) instead of PVdF by casting as an aqueous slurry onto Ti foil with a doctor blade and drying at 80 °C under vacuum. The mass loading of LTO, LFP, LMO, NMC, or activated carbon in the electrodes was in the range of 1.6–2.1, 1.4–1.9, 1.6–2.1, 1.0–1.3 and 2.0–2.5 mg cm⁻², respectively.

2.3 Electrochemical characterization

Linear sweep voltammetry (LSV) was conducted using a three-electrode cell (SB1A, EC FRONTIER) with Pt foil as the working electrode, an activated carbon counter electrode, and an Ag⁺/Ag reference electrode. The reference electrode consisted of a fritted capillary containing 3 mM silver trifluoromethanesulfonate in LiFSA : urea (1 : 4) and Ag wire (= 0.32 V vs. Fc⁺/Fc). The reference electrode potential vs. Li⁺/Li was defined using the redox potential of LTO electrodes (1.55 V vs. Li⁺/Li = −1.75 V vs. Ag⁺/Ag). All LSV were conducted at 0.5 mV s⁻¹. Galvanostatic charge-discharge tests of the negative and positive electrodes were conducted using a three-electrode cell (SB9, EC FRONTIER). In the cell, an activated carbon counter electrode, Ag⁺/Ag reference electrode, and glass fiber separator (GB-100R, Advantec) were used. Measurements were conducted using a VMP3 (Biologic) instrument at 25 ± 5 °C.

2.4 Electrolyte and electrode characterization

The glass transition temperature and melting point of our LiFSA : urea derivative mixtures were analyzed using differential scanning calorimetry (DSC, NETZSCH DSC3500 Sirius) using a temperature scanning rate of 10 °C min⁻¹. Solution structures of the electrolytes were analyzed using Raman spectroscopy (Raman-11, Nanophoton) with a 532 nm laser. Raman spectra were calibrated using the peak originating from silicon at 520 cm⁻¹.²⁸ FT-IR measurements were conducted using Alpha II (Bruker) with the attenuated total reflection (ATR) method inside a glove box to avoid air exposure. The ionic conductivities of the electrolyte solutions were measured using an ionic conductivity meter (CM-41X, TOA DKK). The viscosities of the electrolyte solutions were measured using a rolling ball viscometer (Lovis 2000 M/ME, Anton Paar) at 25 °C.

X-ray photoelectron spectroscopy (XPS, VersaProbe II, ULVAC-PHI) of LTO electrodes was acquired using Al Kα radiation (1486.6 eV) at the National Institute for Materials Science (NIMS) in Japan. The hard X-ray photoelectron spectroscopy (HAXPES) spectra LTO electrodes were acquired using a high excitation energy (7939 eV) and a photoelectron energy analyzer (R-4000, Scienta Omicron) at BL46XU, SPring-8, Japan. The photoelectron detection angle and pass energy of the analyzer were 80° and 200 eV, respectively. Before XPS and HAXPES measurements, the battery electrodes were removed from the three-electrode cells after 50 cycles, rinsed twice with 500 µL DME in an Ar-filled glove box, dried at 25 °C ± 5 °C, and transferred using a transfer vessel to avoid air exposure. The detailed setup and conditions of the HAXPES measurements have been described in our previous reports.^29,30 The binding energy of the obtained spectrum was calibrated using the binding energy of the sp² carbon at 284.3 eV.

3. Results and Discussion

3.1 Structure, physical, and electrochemical properties

We started by systematically investigating DEEs using urea and its methyl derivatives, 1-methylurea, 1,1-dimethylurea (1,1-DMU), and 1,3-dimethylurea (1,3-DMU), and tetramethylurea (TMU), combined with LiFSA (Table 1). As trimethylurea was commercially-available but costly, we did not examine it here. In this way, the hydrogen content of the urea structure was adjusted to impact the hydrogen bond donor behavior. While TMU is liquid solvent at room temperature, the methyl- and dimethyl-urea are solids at room temperature and have melting points close to 100 °C. We visually checked for the formation of liquid or solid-state binary mixtures when the LiFSA and urea derivatives were thoroughly mixed at a molar ratio of 1 : 2 or 1 : 4 (LiFSA : urea derivative). As shown in Fig. 1, the two solid precursors were mixed by stirring at 25 ± 5 °C for 10 hours in an Ar-filled glove box to produce the eutectic melt. We found that LiFSA : urea (1 : 4), LiFSA : 1,3-DMU (1 : 2), and LiFSA : 1,3-DMU (1 : 4) yielded completely liquidified mixtures at 25 °C (Table 1). However, TMU, shown at the bottom in Table 1, is an organic solvent that is liquid at room temperature, and we prepared it for comparison with the deep eutectic electrolytes. The LiFSA concentrations of these electrolytes are 3.64 M (= mol L⁻¹), 3.92 M, 2.37 M, respectively. On the other hand, urea (1 : 2) and both mixtures containing 1,1-DMU or 1-methylurea showed residual solid phases and did not form a deep eutectic melt. This suggests that urea derivatives, and perhaps other donor species, having symmetrical molecular structures will be more likely to form DEEs for application at room temperature (RT), perhaps due to stronger coordination with Li⁺. Detailed interactions are currently under investigation in our laboratory.

Table 1. Urea derivatives for preparing DEEs with LiFSA.

Figure 1.

Photographs of a DEE prepared from starting materials, LiFSA and urea powders. Final liquid product shown on right for LiFSA : urea (1 : 4).

Next, we measured the ionic conductivity and viscosity of our eutectic liquids. As shown in Table 2, all of the melts are found to have ionic conductivities in the range between 1 and 0.1 mS cm⁻¹ with relatively high viscosities compared with typical carbonate-ester battery electrolytes.³¹ In detail, LiFSA : urea (1 : 4) showed the lowest viscosity (280.5 mPa s) among the 1 : 4 DEEs, while higher LiFSA content in LiFSA : 1,3-DMU (1 : 2) showed the highest viscosity of 2533 mPa s. Further, LiFSA : urea (1 : 4) showed the highest conductivity of 1.03 mS cm⁻¹ among them, though all were relatively low (1 to 2 orders of magnitude lower than typical organic electrolytes). We further note that the LiFSA : TMU (1 : 4) electrolyte solution showed similar conductivities and viscosities to LiFSA : urea. A Walden plot constructed from these results (Fig. S1) shows urea and 1,3-DMU DEEs located near the Angell “ideal” line.^4,32 Typically, ionic liquids are described as “poorionic” when falling below this line, so, our DEEs show reasonable conductivities given their high viscosities (Fig. S1).^33–35 In LiFSA : 1,3-DMU (1 : 2), there are more AGG than LiFSA : urea (1 : 4), which is the reason for the low ionic conductivity due to the fewer Li⁺ ion carriers. But in the case of LiFSA : 1,3-DMU (1 : 4), it still showed a higher viscosity than LiFSA : urea (1 : 4) due to its bulkier methyl-groups. As with other electrolyte systems, the conductivities and viscosities showed strong dependences on temperature increasing by ∼10-fold at 60 °C (Table S1). The differences between urea and 1,3-DMU DEEs clearly indicate some level of tunability based on the donor/acceptor choice and overall molecular-scale structures.

Table 2. Ionic conductivity and viscosity of our electrolytes at 25 °C.

	Ionic conductivity /mS cm−1	Viscosity/mPa s
Urea (1 : 4)	1.03	280.5
1,3-DMU (1 : 2)	0.114	2533
1,3-DMU (1 : 4)	0.247	703.1
TMU (1 : 2)	0.346	543.3
TMU (1 : 4)	1.48	55.47

Because electrolyte flammability is a key factor for the safety issue of non-aqueous batteries, we conducted a flammability test and compared with a conventional carbonate electrolyte (1.0 M LiPF₆ EC : DMC) and TMU electrolyte (Fig. 2). For this test, a non-flammable glass fiber separator was soaked in each electrolyte (∼500 µL), then, excess electrolyte was removed, and it was placed into an open flame for 10 seconds. In the case of 1.0 M LiPF₆ EC : DMC and LiFSA : TMU (1 : 4) solutions, the electrolyte burned for several seconds after ignition engulfing the glass fiber separator in flame, which is because EC : DMC and TMU are flammable. On the other hand, urea itself is non-flammable, and indeed our DEEs using LiFSA with urea did not catch on fire. Likewise, LiFSA : 1,3-DMU (1 : 4) did not ignite in our flammability test. Because pure LiFSA, urea and 1,3-DMU are solid at room temperature, their vapor pressure should be negligible which is the main reason for their non-flammable property. This improved safety continues to be a key aspect of DEEs compared to conventional organic electrolytes.

Figure 2.

Flammability test of (a) 1.0 M LiPF₆ EC : DMC (b) LiFSA : TMU (1 : 4) (c) LiFSA : urea (1 : 4) (d) LiFSA : 1,3-DMU (1 : 4).

We further investigated the thermal properties of these DEEs using DSC (Figs. 3 and S2). Starting from 25 °C, the temperature was decreased to −100 °C, then heated to 140 °C, and again decreased back to −100 °C in a cycle. (Fig. S3). In Fig. 3, we have plotted the second heating cycle (from −100 °C to 140 °C) of each DEE. In the case of LiFSA : urea (1 : 4), LiFSA : 1,3-DMU (1 : 2), LiFSA : 1,3-DMU (1 : 4), a glass transition temperature (T_g) was observed around −50 °C (Fig. 3) with no other significant peaks at higher temperatures. This indicates the liquid also can be maintained at a temperature far below RT with no notable phase separation nor precipitation during heating and cooling. This is an important aspect for battery electrolytes as freezing and phase separation of the electrolyte may negatively impact performance. A similar result was observed in LiFSA : TMU (1 : 2), but in the case of the 1 : 4 mixture there was an observed endothermic peak indicating T_m, as reported for other concentrated electrolytes.³⁶ We observed similar results for LiFSA/methylurea mixtures (Fig. S2). The T_g indicates the formation of a super cooling liquid, while the additional T_m peak was observed at ca. 51 °C for LiFSA : methylurea (1 : 4) and it is notable that the T_m is depressed compared to the original solid components. Likewise, for 1,1-DMU, which was solid at RT, we again observed T_m peaks at depressed temperatures, and the emergence of a T_g transition as the urea concentration was decreased. As seen, the ratio and nature of the urea derivative strongly impact the thermal properties of the DEE.

Figure 3.

DSC curves during heating of our DEEs compared with concentrated electrolytes using TMU.

Focusing only on the urea derivatives that formed DEEs, we further investigated the molecular interactions and solution structure using Raman spectroscopy. We started by investigating LiFSA+1,3-DMU DEEs which could be prepared using a wide compositional ratio of 1 : 1.3–1 : 5. Focusing on the S-N-S stretching mode (700–750 cm⁻¹) and S=O stretching mode (1200–1220 cm⁻¹) peaks of FSA⁻, we observed a gradual shift in the peak positions and widths with changing DEE ratio (Figs. 4a and 4b).^37,38 The peaks are normalized by the area of the S-N-S band from FSA⁻. As the concentration of LiFSA increased (and 1,3-DMU decreased), both the S-N-S and S=O peaks shifted to higher wavenumbers and the peaks broadened. This indicates the formation of aggregate structures (AGG) with FSA⁻ coordinated with two or more Li⁺ ions, as observed for other highly concentrated electrolytes.⁸ The spectral regions related to the C=O stretching mode and N-H stretching mode (anti-symmetric and symmetric) moieties of 1,3-DMU (Figs. 4c and 4d, respectively) were also impacted by changes in the DEE ratio.^39–41 In this case, the peaks indicative of free and coordinating 1,3-DMU both increase with decreasing LiFSA concentration. At high 1,3-DMU ratio, e.g. 1 : 5, the liquid structure would contain more free 1,3-DMU molecules without direct interaction of the Li⁺ ions or FSA⁻ anions. Similar peaks and broadening (Fig. 4e, S4) were observed for the LiFSA : urea DEE (1 : 4), though we could not easily vary the LiFSA concentration for this DEE because the composition range of liquid is narrow as 1 : 3.8–1 : 4.2. Comparing the S=O-derived peaks of LiFSA : urea (1 : 4) and LiFSA : 1,3-DMU (1 : 2), we find that the LiFSA : 1,3-DMU (1 : 2) contains more aggregated species. However, LiFSA : urea shows more AGG compared with LiFSA : 1,3-DMU at the same molar ratio, 1 : 4. Analysis using FT-IR (Fig. S5) indicated the N-H stretching peaks were shifted to higher wavenumbers after DEE formation, confirming increased interaction between N-H and FSA⁻.

Figure 4.

Structure analysis of DEEs. Raman spectra of LiFSA+1,3-DMU (1 : 1.3–1 : 5) for (a) S-N-S stretching mode and (b) S=O stretching mode regions. Raman spectra of LiFSA+1,3-DMU for (c) C=O stretching mode and (d) N-H stretching mode regions. (e) Comparison of the S=O peak for LiFSA : urea (1 : 4) and LiFSA : 1,3-DMU (1 : 2, 1 : 4).

After characterizing the structural and physical properties, we examined the electrochemical stability of our DEEs. As shown in Fig. 5, we used LSV at a Pt foil electrode to evaluate the cathodic and anodic stability of each DEE.¹⁸ We defined the decomposition potential as the point at which a current density of 0.05 mA cm⁻² was observed. The 1 : 4 DEEs for LiFSA : urea and LiFSA : 1,3-DMU showed electrochemically-stable potential windows of ∼3.55 V and 5.69 V, respectively. For LiFSA : TMU (1 : 4), the stability window was 3.68 V, but became smaller with increasing LiFSA. On the low potential side, we noted a peak appearing around −2 V vs. Ag⁺/Ag, which may be reversible hydrogen adsorption as protonic electrolytes are active on Pt foil because the quasi-reversible property of the adsorption and desorption were observed even after further cycling in this region (Fig. S6), which we may be due to adsorption/desorption of protons, urea, or a Li-urea complex.⁴² This peak was strongly observed on Pt foil, but the reactions in this potential region were greatly reduced when switching to Al foil (Fig. S7), which is generally known to be incactive for the hydrogen adsorption. Still, 1,3-DMU at a 1 : 2 molar ratio exhibited an exceptionally wide potential window of 6.42 V as shown in Fig. 5, we noted that the cathodic limit is lower than the E° (Li⁺/Li) referenced to our Ag⁺/Ag reference electrode. It is likely that the electrocatalytic behavior of the Pt electrode induces the formation of a highly resistive layer through cathodic decomposition of electrolyte, resulting in the formation of a passivation layer which does not show any notable Li⁺ ion conductivity. This window is exceptional for attaining high-voltage DEE-based batteries and protonicity shows no ill effects.

Figure 5.

LSV curves on Pt foil for LiFSA : urea, LiFSA : 1,3-DMU and LiFSA : TMU. The decomposition potential is defined as the potential at which faradic current reached at −0.05 mA cm⁻².

3.2 Cycling performance with LIB electrodes

Aside from the electrolyte properties, we also wanted to investigate how battery electrodes perform within our Li-conducting DEEs. We started by testing the impact of the LiFSA : 1,3-DMU ratio on the charge-discharge behavior of LTO. Due to its negligible volume change during charge/discharge and high capacity retention, LTO have been actively researched as a potential negative electrode for LIBs.^43,44 As shown in Figs. 6a and 6b, reversible charge-discharge curves were observed for LTO in all of the 1,3-DMU DEEs with initial discharge capacities of 166, 161, and 130 mAh g⁻¹ for 1 : 2, 1 : 3, and 1 : 4 ratios, respectively. These are close to the theoretical capacity for LTO of 175 mAh g⁻¹.⁴³ As the concentration of LiFSA increased, we observed higher initial capacities, improved CEs, and improved capacity retention while cycling at 1 C (175 mA g⁻¹). We found the capacity retention was highest at 87 % after 50 cycles for the LiFSA : 1,3-DMU (1 : 2) DEE. Though the higher LiFSA concentrations led to aggregation, higher viscosities, and lower conductivities, it may play a role in stabilizing the electrode-electrolyte interface as observed in the LSV.⁸ Several reports have discussed the electrochemical reduction of FSA⁻ and resultant formation of solid-electrolyte interphase (SEI) structures. Due to its relatively positive redox potential, LTO does not typically need passivation by SEI in common carbonate electrolytes.⁴⁵ However, the SEI would be necessary in protic solvent electrolytes such as aqueous electrolytes and DEEs, and the highly concentrated electrolyte environment strongly impacts the activity of the anion leading to a shift in the LUMO and more positive reduction potentials to form SEI.⁸ Our group has observed the formation of SEI-type structures at composite electrodes under high concentrations of FSA⁻ as high as ∼1.7 V vs. Li⁺/Li (∼ −1.3 V vs. Ag⁺/Ag).⁴⁶

Figure 6.

Electrochemical perfromance of LTO negative electrodes. (a) Charge-discharge curves and (b) cycle characteristics at 1 C at room temperature (25–30 °C) of LTO with LiFSA : 1,3-DMU = 1 : 2, 1 : 3, 1 : 4 to compare the deep eutectic electrolyte concentration dependence. (c) Charge-discharge curves and (d) cycle characteristics at 1 C at room temperature of LTO in LiFSA : urea (1 : 4), LiFSA : 1,3-DMU (1 : 2), LiFSA : TMU (1 : 2) to compare differences due to urea derivatives.

When testing the response of LTO in the LiFSA : urea (1 : 4) DEE, we observed similar capacities to LiFSA : 1,3-DMU (1 : 2) (Figs. 6c and 6d) and even higher capacity retention at 88 % after 50 cycles. At the same time, the CEs remained quite low throughout cycling, which may be due to continuous electrolyte decomposition. The polarizations of these charge-discharge curves are due to the high viscosity and high rate (As shown in Fig. S8, the lower the rate we chose, the smaller the polarization is.). Rapid fading was observed with LiFSA : TMU (1 : 2), which also showed very low CEs throughout cycling, again suggesting some decomposition reactions involving the electrolyte. Overall, our results point to electrode-electrolyte chemistry as a key component in determining the cycling performance. The wide potential window may also be due to the excellent SEI, and the reactions on the Pt foil and negative or positive electrode are different. Furthermore, we found the binder choice to be critical with PVdF performing better than CMC (Fig. S9), as the CMC was susceptible to dissolution within the DEE.

We further investigated the charge/discharge behavior at slightly elevated temperatures of 35 °C and 40 °C for the LiFSA : 1,3-DMU (1 : 2) (Fig. S10a). At these elevated temperatures, the viscosities and ionic conductivities should be somewhat improved, and we did see some improved capacity retention at 35 °C compared to 25 °C. However, the performance at 40 °C was poorer with low first-cycle CEs and poor capacity retention. We suspected this could be due to the dissolution of the SEI-type structure or the electrode.⁴⁷ Indeed, for another cell we increased the temperature to 40 °C after 50 cycles at room temperature (Fig. S10b), and the discharge capacity remained stable. The development of interphase structures is a complex process, but our results indicate a stable SEI forming at room temperature. As with other electrolytes,^48,49 the DEE structure (e.g. derivative, anion), component concentration, and temperature strongly impact battery electrode charge/discharge performance.

Next, we evaluated the charge-discharge behavior for common positive electrodes that are utilized in LIBs. We focused only on the LiFSA : 1,3-DMU (1 : 2) and LiFSA : urea (1 : 4) and followed similar methods to how the LTO electrodes were tested. Within both DEEs, the charge-discharge of LFP showed excellent performance with high capacity retention and CEs near 100 % (Figs. 7a and 7b). The plateau potential of LFP does not approach the decomposition limit of either DEE, which likely encourages good performance. The LiFSA : 1,3-DMU (1 : 2) electrolyte exhibited larger polarization due to its higher viscosity, while simultaneously showing slightly improved CEs.

Figure 7.

Cycling of positive electrode materials. (a) Charge-discharge curves and (b) extracted capacities and efficiencies for LFP at 0.1 C at room temperature. (c) Charge-discharge curves and (d) cycle comparison for LMO at 0.2 C at 35 °C. The DEE were LiFSA : urea (1 : 4) or LiFSA : 1,3-DMU (1 : 2).

Moving to a higher voltage cathode, LMO, we also observed relatively stable charge-discharge cycling with LiFSA : urea (1 : 4) for 50 cycles at 35 °C with a capacity retention of 70.9 % (Figs. 7c and 7d). However, in the case of LiFSA : 1,3-DMU (1 : 2), the capacity fading was severe with no reversible charge-discharge capacities observed after 20 cycles. The XRD pattern of the LMO electrode after cycling (Fig. S11) showed no change in the crystal structure, so we believe the fading is related electrolyte decomposition at the electrode surface. Another positive electrode, LiNi_1/3Co_1/3Mn_1/3O₂ (NMC111) showed poor cyclability with LiFSA : urea and LiFSA : 1,3-DMU DEEs, despite using a lower cutoff potential of 4.0 V vs. Li⁺/Li (Fig. S12). Considering the varied redox potential of the explored materials with each DEE, our results suggest the performance limitations are linked to electrolyte decomposition, surface chemistry, viscocity and the redox potential of the energy storage reactions. Also it has been reported that protons can insert into the layered oxides,^50,51 so the protons in the DEEs may negatively impact the electrode performance but the crystal structure was not impacted. We speculate the layered structures and their insertion mechanism may show some limitation during desolvation or even some insertion/decomposition of the solvent molecules.

3.3 Surface characterization

Beyond electrolyte properties and electrochemical performance, we further investigated the interfacial chemistry occurring after charge/discharge. Generally, interphase structures are very important for operation of low-voltage negative electrodes in LIBs and other batteries.⁵² Though LTO does not typically form an SEI in conventional EC-based electrolytes as discussed earlier,⁴⁵ the electrolyte environment heavily impacts the redox potentials at which electrolyte decomposition can occur.⁵³ After cycling of the LTO electrodes for 50 charge/discharge cycles, we evaluated their interfacial properties using X-ray photoelectron spectroscopy (XPS) and SEM images (Fig. S13). The SEM images confirm that the electrode surface before and after cycles in DEEs was not remarkably different, but the surface layer is different from the XPS data. All XPS data were normalized based on the area of Ti 2p for comparison. For the F 1s region (Figs. 8a–8c), we found similar surface composition for LiFSA : urea (1 : 4) and LiFSA : 1,3-DMU (1 : 4) DEE. In contrast, LiFSA : 1,3-DMU (1 : 2) (Fig. 8b) showed a much more significant peak at ∼686 eV, indicative of LiF formation. Furthermore, in the XPS results of the surface of the LTO electrode before cycling soaked in LiFSA : 1,3-DMU (1 : 2) (Fig. S8), the Li-F peak cannot be observed, so we identify it is a film component derived from the electrolyte. LiF is known as a common SEI component and known to contribute to SEI stability due to its low solubility and high Young’s modulus.^54,55 For the N 1s (Fig. 8d) and C 1s (Fig. 8e) regions, LiFSA : urea (1 : 4) and LiFSA : 1,3-DMU (1 : 4) again showed similar peaks and intensities.^19,55–57 For LiFSA : 1,3-DMU (1 : 2), the N 1s spectrum shows more significant S-N and C-N species compared with the other two electrolytes, possibly derived from the FSA⁻ anion. There are also indications of more significant carbon content at the surface (e.g. C-N and C=O), which may involve decomposition reactions with 1,3-DMU.

Figure 8.

Surface characterization of LTO electordes. (a–c) XPS spectra of F 1s, (d) N 1s, and (e) C 1s regions for LTO negative electrode after 50 cycles. Normalization was performed using areas of the Ti 2s peak.

We conducted further analyses using hard X-ray photoelectron spectroscopy (HAXPES) to probe deeper regions of the electrode surface.^58,59 As shown in Fig. 9a, LiF-rich layer derived from LiFSA is formed when LiFSA : 1,3-DMU (1 : 2) electrolyte is used in agreement with our XPS analysis. Even in the case of LiFSA : 1,3-DMU (1 : 4), we could find indications of LiF deposited on the electrode. However, for LiFSA : urea (1 : 4), the LiF peak was hard to deconvolute, again agreeing with the XPS results. Looking at the C 1s region in Fig. 9b, we mainly observed the sp² carbon peak with little indication of urea derivative-derived species, such as C-O and C-N. XPS, which observes shallow areas of the surface (∼4 nm), showed substances derived from organic compounds, but HAXPES, which observes deeper areas (∼10 nm), accounts for a small ratio. Overall, this suggests a layered structure for the SEI with organic content mainly observed at outer regions of the SEI (Figs. 9c and 9d), which shows similar trend of the SEI formed in typical carbonate ester electrolyte solution.²⁹ The passivation behavior of low-potential active materials plays critical role for the reversibility, cyclability, and kinetics of lithium insertion and extraction reaction. We believe that further design of DEE itself and interfacial properties of the battery electrodes will improve the battery performance and operation.

Figure 9.

HAXPES analysis of cycled LTO electrodes. HAXPES results for (a) F 1s (b) C 1s regions on LTO negative electrodes after 50 cycles. Normalization was performed using areas of the Ti2p peak. Proposed structure of passive layer in (c) LiFSA : urea (1 : 4) (d) LiFSA : 1,3-DMU (1 : 2). C-F in the illustration indicates C-F species.

4. Conclusion

Throughout this work, we explored and developed new Li⁺-conducting melts, i.e. deep eutectic electrolytes composed of LiFSA with urea derivatives. By changing the structure of the methylurea derivative, we found a notable impact on the propensity of the DEEs to become a liquid electrolyte at room temperature as well as the electrochemical performance. Urea, a promising low-cost DEE components, showed a limited concentration range when forming DEEs. Raman spectroscopy indicated more significant aggregation in this electrolyte compared with 1,3-DMU when prepared at the same molar concentration. At the same time, urea showed higher ionic conductivity and lower viscosity than the 1,3-DMU DEEs. When evaluating the electrolyte stability, we found LiFSA : 1,3-DMU (1 : 2) to exhibit the widest potential window of 6.42 V. However, this did not directly translate to improved electrode performance, as similar or better results were observed for LiFSA : urea (1 : 4) with both positive and negative electrodes. Furthermore, we found quite different results when changing the positive electrode material with stable behavior for LFP, while Li-Mn spinel and NMC111 showed rapid capacity fading and poor performance. Lastly, we found that the DEE concentration impacted the surface species/interphase structure, i.e. DEE containing a high LiFSA content shows more substantial FSA-derived components in the SEI. Overall, our results indicate interesting characteristics for DEEs usage in batteries where the DEE structure strongly impacts stability and performance, especially LiFSA : 1,3-DMU (1 : 2) can form the stable Li-F rich SEI. Also we should keep on studying to increase the ionic conductivity of the DEE. We believe this work will encourage further research to understand the complex solvent properties encountered in DEE, and their development toward practical application of safe, inexpensive rechargeable batteries.

Acknowledgment

XPS measurements were carried out at National Institute for Materials Science (NIMS) Battery Research Platform. Synchrotron radiation experiments were performed at the BL46XU of SPring-8 as the Priority Research Proposal (priority field: Industrial Application) with the approval of the Japan Synchrotron Radiation Research Institute (JASRI) (Proposal No. 2023B1630). The authors thank Dr. Satoshi Yasuno for experimental assistance. This study was partially funded by the JST through CREST (Grant No. JPMJCR21O6), ASPIRE (JPMJAP2313), and GteX (JPMJGX23S4), the MEXT Program: Data Creation and Utilization Type Materials Research and Development Project (JPMXP1122712807), and JSPS KAKENHI (23K20280, 24H00042, 23K26386, 23K13829, 22K14772).

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.28252493.

• 1) Deep eutectic electrolytes (DEEs) are attracting increasing attention as liquid-state electrolytes for secondary batteries because they are potentially low cost, display low flammability, and are environmentally friendly. However, to date limited DEEs have been developed and explored for lithium-ion battery (LIB) applications, with most reports showing unsatisfactory capacity retention, a narrow potential window for battery operation, and an unstable solid electrolyte interphase (SEI) layer leading. Herein, we develop DEEs based on lithium bis(fluorosulfonyl)amide, LiFSA, and a series of urea derivatives as Li ion-conducting DEEs. Despite similar structures for the urea derivatives, i.e. methylated urea, we found that 1,3-dimethylurea (1,3-DMU) could form Li ion-concentrated DEEs across a wide range of LiFSA : 1,3-DMU ratio, while the LiFSA : urea DEE was liquid only in a limited range of molar ratios, i.e. LiFSA : urea close to 1 : 4 (mol/mol). By examining the electrolyte structure via Raman spectroscopy, we observed increased aggregation for DEE with higher LiFSA concentrations. We further confirmed non-flammability and electrochemical stability among the DEEs with potential windows ranging from ∼3.35 V for LiFSA : urea (1 : 4) to an impressive 6.42 V for LiFSA : 1,3-DMU (1 : 2) at a Pt foil electrode. During charge-discharge of Li₄Ti₅O₁₂ (LTO) electrodes, we observed good capacities and retention for the LiFSA : urea (1 : 4) and LiFSA : 1,3-DMU (1 : 2) DEEs. High Coulombic efficiencies (CEs) were achieved in the LiFSA : 1,3-DMU (1 : 2) DEE with its high LiFSA content that led to more substantial FSA-derived components in the SEI structures after cycling. We further tested positive electrode materials, including LiFePO₄ that showed excellent capacity retention and CEs near 100 % across 50 cycles. In all, we find that the dimethylurea-based DEEs show an opportunity for non-flammable and high-voltage Li batteries.

CRediT Authorship Contribution Statement

Nanako Ito: Data curation (Lead), Formal analysis (Lead), Writing – original draft (Lead)

Tomooki Hosaka: Conceptualization (Lead), Investigation (Supporting), Writing – review & editing (Lead)

Ryoichi Tatara: Investigation (Supporting), Methodology (Equal), Writing – review & editing (Equal)

Zachary T. Gossage: Formal analysis (Supporting), Investigation (Supporting), Writing – original draft (Equal)

Shinichi Komaba: Supervision (Lead), Writing – review & editing (Supporting)

Conflict of Interest

The authors declare no conflict of interest in the manuscript.

Funding

JST, MEXT: JPMJCR21O6, JPMJAP2313, JPMJGX23S4, JPMXP1122712807,

JSPS: 23K20280, 24H00042, 23K26386, 23K13829, 22K14772

Footnotes

N. Ito: ECSJ Student Member

T. Hosaka, R. Tatara, and Z. T. Gossage: ECSJ Active Members

S. Komaba: ECSJ Fellow

References

• 1)  M. Li, J. Lu, Z. Chen, and K. Amine, Adv. Mater., 30, 1800561 (2018).
• 2)  K. Xu, Chem. Rev., 104, 4303 (2004).
• 3)  D.-H. Liu, Z. Bai, M. Li, A. Yu, D. Luo, W. Liu, L. Yang, J. Lu, K. Amine, and Z. Chen, Chem. Soc. Rev., 49, 5407 (2020).
• 4)  X. Feng, M. Ouyang, X. Liu, L. Lu, Y. Xia, and X. He, Energy Storage Mater., 10, 246 (2018).
• 5)  X. Tian, Y. Yi, B. Fang, P. Yang, T. Wang, P. Liu, L. Qu, M. Li, and S. Zhang, Chem. Mater., 32, 9821 (2020).
• 6)  Q. Zheng, Y. Yamada, R. Shang, S. Ko, Y.-Y. Lee, K. Kim, E. Nakamura, and A. Yamada, Nat. Energy, 5, 291 (2020).
• 7)  J. Wang, Y. Yamada, K. Sodeyama, E. Watanabe, K. Takada, Y. Tateyama, and A. Yamada, Nat. Energy, 3, 22 (2018).
• 8)  Y. Yamada, K. Furukawa, K. Sodeyama, K. Kikuchi, M. Yaegashi, Y. Tateyama, and A. Yamada, J. Am. Chem. Soc., 136, 5039 (2014).
• 9)  Y. Yamada, J. Wang, S. Ko, E. Watanabe, and A. Yamada, Nat. Energy, 4, 269 (2019).
• 10)  Y. Yamada, K. Usui, K. Sodeyama, S. Ko, Y. Tateyama, and A. Yamada, Nat. Energy, 1, 16129 (2016).
• 11)  T. Hosaka, R. Takahashi, K. Kubota, R. Tatara, Y. Matsuda, K. Ida, K. Kuba, and S. Komaba, J. Power Sources, 548, 232096 (2022).
• 12)  M. Hamada, R. Tatara, K. Kubota, S. Kumakura, and S. Komaba, ACS Energy Lett., 7, 2244 (2022).
• 13)  Q. Zhao, S. Stalin, C.-Z. Zhao, and L. A. Archer, Nat. Rev. Mater., 5, 229 (2020).
• 14)  Q. Zhang, K. D. O. Vigier, S. Royer, and F. Jérôme, Chem. Soc. Rev., 41, 7108 (2012).
• 15)  A. Boisset, S. Menne, J. Jacquemin, A. Balducci, and M. Anouti, Phys. Chem. Chem. Phys., 15, 20054 (2013).
• 16)  D. De Sloovere, D. E. P. Vanpoucke, A. Paulus, B. Joos, L. Calvi, T. Vranken, G. Reekmans, P. Adriaensens, N. Eshraghi, A. Mahmoud, F. Boschini, M. Safari, M. K. Van Bael, and A. Hardy, Adv. Energy Sustain. Res., 3, 2100159 (2022).
• 17)  E. L. Smith, A. P. Abbott, and K. S. Ryder, Chem. Rev., 114, 11060 (2014).
• 18)  N. Z. Hardin, Z. Duca, A. Imel, and P. A. Ward, ChemElectroChem, 9, e202200628 (2022).
• 19)  J. Zhang, H. Wu, X. Du, H. Zhang, L. Huang, F. Sun, T. Liu, S. Tian, L. Zhou, S. Hu, Z. Yuan, B. Zhang, J. Zhang, and G. Cui, Adv. Energy Mater., 13, 2202529 (2023).
• 20)  A.-S. Kelchtermans, D. De Sloovere, J. Mercken, T. Vranken, G. Mangione, B. Joos, W. Vercruysse, D. Vandamme, H. Hamed, M. Safari, E. Derveaux, P. Adriaensens, M. K. Van Bael, and A. Hardy, ACS Omega, 9, 42343 (2024).
• 21)  I. Sumarlan, A. Kunverji, A. J. Lucio, A. R. Hillman, and K. S. Ryder, J. Phys. Chem. C, 127, 18891 (2023).
• 22)  S. Azmi, M. F. Koudahi, and E. Frackowiak, Energy Environ. Sci., 15, 1156 (2022).
• 23)  L. Geng, X. Wang, K. Han, P. Hu, L. Zhou, Y. Zhao, W. Luo, and L. Mai, ACS Energy Lett., 7, 247 (2022).
• 24)  M. Angell, C. Pan, Y. Rong, C. Yuan, M. Lin, B. Hwang, and H. Daia, Proc. Natl. Acad. Sci. U.S.A., 114, 834 (2017).
• 25)  J. Qian, W. A. Henderson, W. Xu, P. Bhattacharya, M. Engelhard, O. Borodin, and J.-G. Zhang, Nat. Commun., 6, 6362 (2015).
• 26)  Z. Song, X. Wang, H. Wu, W. Feng, J. Nie, H. Yu, X. Huang, M. Armand, and H. Zhang, J. Power Sources, 14, 100088 (2022).
• 27)  H. Ogawa and H. Mori, Phys. Chem. Chem. Phys., 22, 8853 (2020).
• 28)  M. Park, C. W. Teng, V. Sakhrani, M. B. McLaurin, R. M. Kolbas, R. C. Sanwal, R. J. Nemanich, J. J. Hren, and J. J. Cuomo, J. Appl. Phys., 89, 1130 (2001).
• 29)  N. Yabuuchi, K. Shimomura, Y. Shimbe, T. Ozeki, J. Y. Son, H. Oji, Y. Katayama, T. Miura, and S. Komaba, Adv. Energy Mater., 1, 759 (2011).
• 30)  M. Dahbi, T. Nakano, N. Yabuuchi, S. Fujimura, K. Chihara, K. Kubota, J.-Y. Son, Y.-T. Cui, H. Oji, and S. Komaba, ChemElectroChem, 3, 1856 (2016).
• 31)  E. R. Logan, E. M. Tonita, K. L. Gering, J. Li, X. Ma, L. Y. Beaulieu, and J. R. Dahn, J. Electrochem. Soc., 165, A21 (2018).
• 32)  W. Xu, E. I. Cooper, and C. A. Angell, J. Phys. Chem. B, 107, 6170 (2003).
• 33)  L. Li, H. Cheng, J. Zhang, Y. Guo, C. Sun, M. Zhou, Q. Li, Z. Ma, and J. Ming, ACS Energy Lett., 8, 1076 (2023).
• 34)  K. R. Harris, J. Phys. Chem. B, 123, 7014 (2019).
• 35)  C. Schreiner, S. Zugmann, R. Hartl, and H. J. Gores, J. Chem. Eng. Data, 55, 1784 (2010).
• 36)  K. Takada, Y. Yamada, E. Watanabe, J. Wang, K. Sodeyama, Y. Tateyama, K. Hirata, T. Kawase, and A. Yamada, ACS Appl. Mater. Interfaces, 9, 33802 (2017).
• 37)  Y. Yamada, M. Yaegashi, T. Abe, and A. Yamada, Chem. Commun., 49, 11194 (2013).
• 38)  K. Fujii, H. Hamano, H. Doi, X. Song, S. Tsuzuki, K. Hayamizu, S. Seki, Y. Kameda, K. Dokko, M. Watanabe, and Y. Umebayashi, J. Phys. Chem. C, 117, 19314 (2013).
• 39)  L. P. Silva, C. F. Araújo, D. O. Abranches, M. Melle-Franco, M. A. R. Martins, M. M. Nolasco, P. J. A. Ribeiro-Claro, S. P. Pinho, and J. A. P. Coutinho, Phys. Chem. Chem. Phys., 21, 18278 (2019).
• 40)  R. L. Frost, J. Kristof, L. Rintoul, and J. T. Kloprogge, Spectrochim. Acta, Part A, 56, 1681 (2000).
• 41)  R. Keuleers, H. O. Desseyn, B. Rousseau, and C. Van Alsenoy, J. Phys. Chem. A, 103, 4621 (1999).
• 42)  M. F. Suárez-Herrera, M. Costa-Figueiredo, and J. M. Feliu, Langmuir, 28, 5286 (2012).
• 43)  T. Ohzuku, A. Ueda, and N. Yamamoto, J. Electrochem. Soc., 142, 1431 (1995).
• 44)  C. Wang, S. Wang, Y.-B. He, L. Tang, C. Han, C. Yang, M. Wagemaker, B. Li, Q.-H. Yang, J.-K. Kim, and F. Kang, Chem. Mater., 27, 5647 (2015).
• 45)  Y.-B. He, M. Liu, Z.-D. Huang, B. Zhang, Y. Yu, B. Li, F. Kang, and J.-K. Kim, J. Power Sources, 239, 269 (2013).
• 46)  Z. T. Gossage, N. Ito, T. Hosaka, R. Tatara, and S. Komaba, Angew. Chem., 135, e202307446 (2023).
• 47)  D. Reber, R. Figi, R.-S. Kühnel, and C. Battaglia, Electrochim. Acta, 321, 134644 (2019).
• 48)  W. Xiong, X. Zhang, Z. Tu, X. Hu, and Y. Wu, Ind. Eng. Chem. Res., 62, 51 (2023).
• 49)  M. E. Di Pietro, K. Goloviznina, A. van den Bruinhorst, G. de Araujo Lima e Souza, M. Costa Gomes, A. A. H. Padua, and A. Mele, ACS Sustainable Chem. Eng., 10, 11835 (2022).
• 50)  J.-Y. Luo, W.-J. Cui, P. He, and Y.-Y. Xia, Nat. Chem., 2, 760 (2010).
• 51)  Y. Wang, J. Luo, C. Wang, and Y. Xia, J. Electrochem. Soc., 153, A1425 (2006).
• 52)  Z. T. Gossage, T. Hosaka, T. Matsuyama, R. Tatara, and S. Komaba, J. Mater. Chem. A, 11, 914 (2023).
• 53)  Y. Zeng, Z. T. Gossage, D. Sarbapalli, J. Hui, and J. Rodríguez-López, ChemElectroChem, 9, e202101445 (2022).
• 54)  J. Tan, J. Matz, P. Dong, J. Shen, and M. Ye, Adv. Energy Mater., 11, 2100046 (2021).
• 55)  Y. Liang, W. Wu, D. Li, H. Wu, C. Gao, Z. Chen, L. Ci, and J. Zhang, Adv. Energy Mater., 12, 2202493 (2022).
• 56)  W. Wu, Y. Liang, D. Li, Y. Bo, D. Wu, L. Ci, M. Li, and J. Zhang, ACS Nano, 16, 14558 (2022).
• 57)  R. R. Solís, M. A. Quintana, G. Blázquez, M. Calero, and M. J. Muñoz-Batista, Catal. Today, 423, 114266 (2023).
• 58)  S. Yamazaki, R. Tatara, H. Mizuta, K. Kawano, S. Yasuno, and S. Komaba, Mater. Adv., 4, 1637 (2023).
• 59)  T. Hosaka, T. Matsuyama, K. Kubota, S. Yasuno, and S. Komaba, ACS Appl. Mater. Interfaces, 12, 34873 (2020).