2025 Volume 93 Issue 6 Pages 063008
An organic electrolyte solution of 1.0 mol dm−3 LiCl/dimethyl sulfoxide (DMSO) was investigated for a fluorine-free battery with an electrode using 2,5-dimethoxy-1,4-benzoquinone (DMBQ) as the n-type active material and styrene butadiene rubber (SBR) as the binder. The DMBQ electrodes showed the initial discharge capacities of 222 mAh g−1 and 164 mAh g−1 for the DMSO solution of the LiCl system and the LiPF6 system, respectively. Moreover, they showed similar working voltages of 2.6–2.7 V, even though their plateau slopes had slight differences. In addition, the chemical bonding and crystallinity changes of DMBQ were confirmed during discharge-charge, suggesting that it reacted electrochemically with Li+ in the LiCl/DMSO solution. The results of cycle properties suggested that a gradual decrease in the discharge capacities would result from denaturation and dissolution of DMBQ during the redox process in the electrolyte solution.
In the quest for safer and more sustainable batteries, it is essential to identify and develop less hazardous substances, such as fluorinated compounds, also known as PFAS.1–3 Lithium secondary batteries are composed of graphite or lithium metal as the negative electrode material, LiCoO2 (LCO), LiNi1−x−yCoxMnyO2 (NCM), or LiFePO4 (LFP) as the positive electrode material, an organic solution of a lithium salt with a fluorine-containing anion group (such as hexafluorophosphate (PF6) or bis(trifluoromethanesulfonyl)imide (TFSI)) as the electrolyte, and a fluorine-containing polymer such as polytetrafluoroethylene (PTFE) or polyvinylidene fluoride (PVDF) as the binder. However, such fluorine-containing materials have been detected not only in communities near battery manufacturing sites but also in all of the locations at which these batteries are disposed, raising concerns about threats to the environment and human health.4 Therefore, the use of a fluorine-free electrolyte, which does not use fluorine compounds, is desired.
While LiClO4 is a potential candidate for a fluorine-free electrolyte, it is an oxidizing agent that can cause fires when in contact with combustible materials, raising safety concerns.5 Therefore, we have focused on LiCl, which is said to have abundant reserves, little risk to the environment, and chemical stability.6,7
The dual-ion lithium secondary battery reported by Wang et al. in 2022 used a LiCl/dimethyl sulfoxide (DMSO) electrolyte solution, and discharge-charge was proceeded by the deposition/dissolution reaction of Li metal in the anode and the insertion/extraction reaction of Cl− in the cathode of poly(butyl viologen dichloride), a p-type organic active material with a quaternary nitrogen (=N+<) moiety.8 However, there are few options of p-type active materials capable of insertion/extraction of Cl−. Therefore, it is necessary to obtain basic knowledge on the application of an n-type cathode active material capable of insertion/extraction of Li+, which is applicable to electrode materials for rocking chair lithium batteries and has been reported in many cases.9,10 In addition, the working voltage of the cathode must be below 3.1 V because of the Cl− → Cl3− (> 3.1 V vs. Li/Li+) side reaction in the LiCl/DMSO electrolyte solution. Transition metal-containing oxides may also be candidates with respect to the working range of cathodes using n-type active materials. However, while the working voltages of 3.8 V vs. Li/Li+ for LCO,11 3.7 V vs. Li/Li+ for NCM,12 and 3.4 V vs. Li/Li+ for LFP13 have been reported, no candidate materials that satisfy the working condition of the LiCl/DMSO electrolyte solution (≤ 3.1 V vs. Li/Li+) have been identified thus far.
On the other hand, organic compounds such as carbonyl compounds, cyano compounds, imine and azo compounds, and organosulfur have been reported to function as n-type active materials in lithium-, sodium-, and magnesium-based batteries.14–18 Among these, the working voltages of the cathode using quinones, which correspond to carbonyl compounds, is roughly 0.5–3 V vs. Li/Li+ as reported by Liang et al.,19 which is considered to match the working conditions of LiCl/DMSO electrolyte solution. In particular, 2,5-dimethoxy-1,4-benzoquinone (DMBQ) with low solubility in organic solvents is considered to react by the mechanism shown in Scheme 1,20 and as the electrode active material in the LiCl/DMSO electrolyte solution. However, as far as we know, the electrolyte solutions used with DMBQ electrodes are metal salts containing perchloric acid or fluorine, such as LiClO4,20,21 NaClO4,21 magnesium bis(trifluoromethanesulfonyl)amide,22 and Mg(TFSI)2-MgCl2,23 and there have been no reports utilizing LiCl.
Redox mechanism of DMBQ in lithium secondary battery systems.
In this study, we evaluated the discharge-charge characteristics of DMBQ in LiCl/DMSO electrolyte solution with the objective of achieving a fluorine-free lithium secondary battery. As a comparison, the same evaluation was performed on an LiPF6/DMSO electrolyte solution. The electrochemical interaction between n-type active material DMBQ and Li+ in the LiCl/DMSO electrolyte solution was investigated by evaluating the changes in the crystal structure and chemical bonding state of the DMBQ electrode during discharge-charge. Finally, by evaluating the stability of the redox cycle, we identified issues for further performance improvement of the LiCl/DMSO system.
The working electrode consists of DMBQ (Tokyo Chemical Industry) as the active material, styrene butadiene rubber (SBR, 50 wt%, MTI Japan) as the binder, carboxymethyl cellulose (CMC, Cellenpia, Nippon Paper Industries) as the separator, acetylene black (AB, Denka Black, Denka Company) as conductive carbon, and aluminum foil as the current collector. First, an aqueous slurry incorporating electrode materials was prepared to be DMBQ : AB : SBR : CMC = 8 : 1 : 0.8 : 0.2 (wt), which was casted to Al foil. The electrodes were then dried at 40 °C for 24 h, cut, and pressed into 16-mm-diameter samples with the DMBQ loading of 2.2–2.6 mg cm−2. To investigate the electrochemical stability of the DMBQ electrode, a carbon-only electrode without DMBQ, AB : SBR : CMC = 5 : 4 : 1 (wt), was also prepared under the same conditions. In this report, SBR rather than PTFE or PVDF was used as the binder of the DMBQ electrode for the construction of a fluorine-free lithium secondary battery. Under the conditions of use of LiCl/DMSO electrolyte solution, degradation due to oxidation of butadiene units in SBR (≥ 4.2 V vs. Li/Li+)24 is not expected to occur. The electrolyte solution was prepared by mixing and stirring anhydrous LiCl (≥ 99.0 %, Sigma-Aldrich) to 1.0 mol dm−3 in DMSO solvent (H2O ≤ 20 ppm, Tomiyama Pure Chemical Industries). A glass fiber membrane (EL-Cell) was used as the separator. For performance comparison, commercially available 1.0 mol dm−3 LiPF6/DMSO (Tomiyama Pure Chemical Industries) was utilized. A metallic lithium foil (200-µm-thick, Honjo Metal) was used as the counter electrode. For cyclic voltammetry and electrochemical impedance measurements, a lithium strip (EL-Cell) was utilized as the reference electrode. All experiments were performed in a dry room kept below the dew point of −60 °C.
2.2 Electrochemical measurementsThe PAT-CELL (EL-Cell) was used as the evaluation cell. The discharge-charge characteristics of the DMBQ electrode were tested with a potentiostat (VMP-3, Biologic) at 50 µA cm−2 in the voltage range from 1.8 to 3.1 V. Discharge rate characteristics were also tested at current densities of 5 µA cm−2, 50 µA cm−2, 150 µA cm−2, 250 µA cm−2, and 500 µA cm−2 in the voltage range from an open circuit voltage to 1.8 V. For electrochemical stability, cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) were evaluated with a potentiostat (VSP-300, Biologic) in the PAT-CELL (EL-Cell) of a three-electrode cell system. For electrodes with and without DMBQ, EIS was measured over a frequency range of 100 mHz to 5 MHz in the condition of open-circuit state after 1st, 10th and 20th cycle during sweeping at 0.5 mV s−1 in CV in the rage of 1.8–3.1 V. All measurements were performed in the dry room.
2.3 Structural analysis of DMBQ electrode during discharge-chargeChanges in crystallinity and chemical bonding of DMBQ electrodes in LiCl/DMSO electrolyte solution during discharge-charge were evaluated. The DMBQ electrode used for the analysis was removed from the evaluation cell after discharge or charge and washed with dimethyl carbonate (DMC). Fourier transform infrared25,26 (FTIR, FT/IR-680Plus, JASCO) recording in the wavenumber range of 1000–2000 cm−1 was utilized to analyze the chemical bonding states of the DMBQ electrode. X-ray diffraction21 (XRD, Ultima IV, Rigaku) was used to analyze the crystal structure of the DMBQ electrode by measuring in the scan range of 2θ = 5 deg–35 deg utilizing CuKα radiation at a tube voltage of 40 kV and tube current of 40 mA.
2.4 Composition analysis of LiCl/DMSO electrolyte solution after discharge-charge cycleTo evaluate the stability of LiCl/DMSO electrolyte solution, compositional changes and decomposition of the electrolyte solution before and after discharge-charge were analyzed. The Li content was measured with atomic absorption spectrometry27 (AAS, ZA3300, Hitachi High-Tech Science), and the Cl content in the solution was measured with ion chromatography28 (IC, IC-2010, Tosoh). In addition, the DMSO content in the solution was measured using gas chromatography with a flame ionization detector29 (GC/FID, GC-2010, Shimadzu) and proton nuclear magnetic resonance30 (1H-NMR, JNM-ECZ600, JEOL) using tetrabromoethane (TBE) as an internal standard.
The discharge-charge curves of the DMBQ electrode in LiCl/DMSO and LiPF6/DMSO electrolyte solutions are shown in Figs. 1a and 2a, respectively. While there were differences in the slope of the plateau in the LiCl/DMSO and LiPF6/DMSO electrolyte solutions, both electrodes were found to have the working voltage of 2.6–2.7 V. The LiCl system and the LiPF6 system exhibited initial discharge capacities of 70 % (222 mAh g−1) and 51 % (164 mAh g−1) of the theoretical value (319 mAh g−1) assuming a two-electron redox reaction of DMBQ, respectively. Similarly, the previously reported dual-ion battery also showed a higher initial discharge capacity for the LiCl system than for the LiPF6 system.8
Electrochemical properties of DMBQ in 1.0 mol dm−3 LiCl/DMSO electrolyte solution. (a) Discharge-charge curves of DMBQ electrode at current density of 50 µA cm−2. (b) Cycle characteristics. (c) Initial discharge curves at various current densities. (d) Plateau in the initial discharge curve (5 µA cm−2). (e) Cyclic voltammogram curve measured with scanning rate of 0.5 mV s−1 and scanning range of 1.8–3.1 V.
Electrochemical properties of DMBQ in 1.0 mol dm−3 LiPF6/DMSO electrolyte solution. (a) Discharge-charge curves of DMBQ electrode at current density of 50 µA cm−2. (b) Cycle characteristics. (c) Initial discharge curves at various current densities. (d) Plateau in the initial discharge curve (5 µA cm−2). (e) Cyclic voltammogram curve measured with scanning rate of 0.5 mV s−1 and scanning range of 1.8–3.1 V.
However, as shown in Figs. 1b and 2b, the discharge-charge cycle characteristics of both the LiCl system and the LiPF6 system were stable at approximately 165 mAh g−1 from the second to the sixth cycles, but the capacities thereafter declined markedly, decreasing to approximately 13 % and 17 % of the initial capacity at the 20th cycle. These results demonstrate that the trend of capacity decrease with discharge-charge cycle was similar for both systems, presumably due to electrochemical dissolution of DMBQ and degradation of the electrolyte.
As shown in Fig. S1 of the Supporting Information, the electrolyte solution removed from the cell after 20th discharge-charge cycles showed a brown coloration, which was presumably caused by degradation of the electrolyte solution with cycling.
The initial discharge curves of the DMBQ electrodes at various current densities are shown in Figs. 1c and 2c. The discharge capacity of both electrolyte systems decreased with increasing current density. Focusing on the current density of 5 µA cm−2 shown in Figs. 1d and 2d, there was a clear difference in the ratio (CV1st : CV2nd) of the capacity of the first stage discharge plateau (CV1st) to that of the second stage discharge plateau (CV2nd) in both electrolyte systems. Since CV1st : CV2nd ≈ 1 : 1 for the LiPF6 system, as in previous reports,21 the electrode reaction on DMBQ was considered to proceed via a scheme in which Li+ inserts sequentially in the radical monoanionic state in the first stage discharge plateau and in the dianionic state in the second plateau, as shown in Scheme 1. In contrast, the LiCl system has CV1st : CV2nd ≈ 35 : 3, suggesting that the scheme of Li+ insertion to DMBQ was different from that of the LiPF6 system. These findings indicate that DMBQ had a reaction mechanism based on a reversible redox reaction in the LiCl system as well as in the LiPF6 system, as the working voltages were almost identical, although there were differences in the trends at the plateau. Figures 1e and 2e show the CVs of the DMBQ electrode in the LiCl/DMSO and LiPF6/DMSO electrolyte solutions, respectively. One sharp reduction current peak around 2.5 V in the LiCl system and two adjacent reduction current peaks around 2.5 V and 2.6 V in the LiPF6 system were consistent with the trend of the discharge-charge measurement results. In addition, the conductivities of 1 mol dm−3 LiCl/DMSO and 1 mol dm−3 LiPF6/DMSO were reported to be 5.6 mS cm−1 and 11.4 mS cm−1,8,31 respectively, but as shown in Figs. 1e and 2e, the peak cathodic current densities were 1.7 mA cm−2 for the LiCl system and 1.0 mA cm−2 for the LiPF6 system. This suggests that strongly solvated Cl− affects the concentration gradient of Li+, resulting in a reaction environment in which excess Li+ is supplied to DMBQ. Then, in the progression of the two-phase coexistence reactions shown in Scheme 1, different phase transitions may overlap due to the rate-determining electron transfer. On the other hand, in the LiPF6 system, Li+ is smoothly supplied in response to electron transfer to DMBQ, so each phase transition is expected to proceed well separated.
Pan et al. reported two-step and one-step discharge curves for the DMBQ electrode in Mg(TFSI)2/DME and Mg(TFSI)2-2MgCl2/DME electrolyte solutions, respectively, which is similar to the finding in this report that the plateau of DMBQ in the electrolyte solution containing Cl− was one-step. From the above, we can consider the difference in discharge-charge behavior between the LiCl and LiPF6 systems to stem from the Li+ concentration in the vicinity of the DMBQ.
3.2 Redox reaction mechanism of DMBQ electrodeTo investigate the redox reaction mechanism of DMBQ during discharge-charge, in the initial discharge-charge test at a current density of 50 µA cm−2 in LiCl/DMSO electrolyte solution, the measurement was stopped at the time of discharge or charge, and the DMBQ electrode was removed for FTIR and XRD measurements, as shown in Fig. 3. The FTIR spectrum and XRD pattern of DMBQ powder are shown in Fig. S2 of the Supporting Information for reference. The stopping voltages were set at the points shown in Fig. 3a.
Various analysis results of discharged/charged DMBQ electrode in 1.0 mol dm−3 LiCl/DMSO electrolyte solution. (a) Initial discharge-charge curve. (b) FTIR spectrum. (c) XRD pattern. Measurement points are at the beginning of the first stage discharge plateau (2.66 V), at the beginning of the second stage discharge plateau (2.64 V), at the end of discharge (1.80 V), at the beginning of the first stage charge plateau (2.76 V), at the beginning of the second stage charge plateau (2.78 V), and at the end of charge (3.10 V).
In the FTIR spectrum of the DMBQ electrode in the LiCl/DMSO electrolyte solution shown in Fig. 3b, the C=O stretching vibration originating from DMBQ (around 1650 cm−1) disappeared and the C=C stretching vibration of the aromatic ring (around 1600 cm−1) gradually shifted toward the low wavenumber side during the discharge process. Then, the C–O−Li+ stretching vibration (around 1200 cm−1) appeared, and these peaks changed to the same spectra as before the start of discharge, confirming the recovery of the pristine state during the charge process. These reversible changes indicate that DMBQ achieved a reversible reduction/oxidation cycle through electrochemical reactions with Li+. In a previous report on 2,5-diamino-1,4-benzoquinone in 1 mol dm−3 LiTFSI in mixed solvent of tetraethylene glycol dimethyl ether and 1,3-dioxolane (1 : 1 in volume ratio), the C=O stretching vibration disappeared and C–O−Li+ stretching vibration appeared with discharge, and the original state was recovered with charge.25 Similar FTIR spectral changes were observed when DMBQ was lithiated in the solid electrolyte Li6PS5Cl to form Li2DMBQ.26 We can therefore assume that, in the insertion/extraction reaction of Li+ in the cathode of DMBQ, an n-type active material with a quinone moiety has progressed.
The XRD pattern shown in Fig. 3c confirms that the amorphization of DMBQ progresses, as its crystalline peaks decay during the discharge process. This is presumably similar to the phenomenon observed when Li+ is accommodated in the crystal lattice of DMBQ, as previously reported.21 On the other hand, the crystalline peaks are observed to return to their original state upon charging, indicating the electrochemical reversibility of the crystalline structure with respect to DMBQ. The same trend was observed for DMBQ in 1 mol dm−3 LiClO4/butyl lactone electrolyte solution, where the crystalline peaks derived from DMBQ disappeared at the end of discharge and returned to their original state at the end of charging.21 From these results, similar to the previous report, we conclude that DMBQ in the electrode is reversible and is able to maintain a stable structure.
The results of FTIR and XRD measurements of the DMBQ electrode in the LiPF6 system under the same conditions as Fig. 3 are shown in Fig. 4. In the FTIR spectrum of the DMBQ electrode in the LiPF6 system shown in Fig. 4b, the chemical bonds that change the stretching vibration were similar to those of the LiCl system, although there were differences in the shift value of the C=C stretching vibration (around 1600 cm−1) toward the lower wavenumber side with discharge. This result is presumably due to the aforementioned difference in plateau as shown in Figs. 1d and 2d. In the XRD pattern shown in Fig. 4c, the trend of crystallinity changes with discharge was similar to that of the LiCl system, although there was a difference in the degree of amorphization progression. This difference might be due to the discharge capacity being 51 % of the theoretical capacity, which means that in this electrolyte solution, a large amount of DMBQ remained in the electrolyte solution whose molecular arrangement had not changed without interacting with Li+. On the other hand, a new peak appeared around 2θ = 8 deg during the charging process, although it disappeared at the end of charging, as in the LiCl system. According to the discharge-charge curves and FT-IR results, the LiPF6 system and the LiCl system have slightly different trends, which may lead to the formation of different intermediate states of DMBQ. These results indicate that the electrode proceeds by a similar redox mechanism in the LiCl system and the LiPF6 system, although there were some differences in the changes in the stretching vibration and the crystallinity of DMBQ during discharge-charge.
Various analysis results of discharged/charged DMBQ electrode in 1.0 mol dm−3 LiPF6/DMSO electrolyte solution. (a) Initial discharge-charge curve. (b) FTIR spectrum. (c) XRD pattern. Measurement points are at the beginning of the first stage discharge plateau (2.69 V), at the beginning of the second stage discharge plateau (2.62 V), at the end of discharge (1.80 V), at the beginning of the first stage charge plateau (2.76 V), at the beginning of the second stage charge plateau (2.78 V), and at the end of charge (3.10 V).
As shown in Figs. 1b and 2b, to investigate the capacity decrease with discharge-charge cycling in both electrolyte solution systems, we analyzed the composition change of the LiCl/DMSO electrolyte solution after discharge-charge, the decomposition products, and the denatured products generated from the electrode materials by dissolution. Specifically, the contents of Li, Cl, and DMSO in the electrolyte solution before and after the discharge-charge cycle were measured using AAS, IC, GC/FID, and 1H-NMR. As shown in Table 1, the molar ratios of Li/Cl in the electrolyte solution obtained from the results of AAS and IC were 1.07 and 1.02 before and after 20 discharge-charge cycles, respectively. The mass percent of DMSO in the electrolyte solution before and after 20 discharge-charge cycles was 89.3 wt% and 88.6 wt% for GC/FID and 92.7 wt% and 91.1 wt% for 1H-NMR, respectively.
| Quantitative item | Analysis method | Composition | ||
|---|---|---|---|---|
| Before | After 20th cycle | |||
| Electrolyte | Li atom (wt%) | Atomic absorption spectrometry | 0.565 | 0.557 |
| Cl atom (wt%) | Ion chromatography | 2.72 | 2.83 | |
| Li/Cl (in molar ratio) | — | 1.07 | 1.02 | |
| Solvent | DMSO (wt%) | Gas chromatography with a flame ionization detector | 89.3* | 88.6* |
| Proton nuclear magnetic resonance | 92.7 | 91.1 | ||
*: Average of two samples.
According to the results of these quantitative evaluations, the denaturation of LiCl and DMSO caused by the composition change and decomposition of the LiCl/DMSO electrolyte solution is considered to be minor after the discharge-charge cycles.
Next, to evaluate the stability of DMBQ in electrodes in the LiCl system during redox cycling, we utilized a three-electrode electrochemical cell with lithium metal as the counter electrode and reference electrode, a DMBQ electrode, and a carbon-only electrode (AB : SBR : CMC = 5 : 4 : 1 (wt)) without DMBQ as the working electrode. The carbon-only electrode without DMBQ had rather small current and no reduction/oxidation peaks, as shown in Fig. 5a, and a smaller impedance change, as shown in Fig. 5b than that of the DMBQ electrode. Since the upper voltage was set to 3.1 V, which was the voltage at which the oxidative decomposition of Cl− → Cl3− did not occur in the LiCl system.8 Therefore, a large decomposition current of the electrolyte solution did not flow, and the resistance change was considered to be suppressed as shown in Fig. 5b.
Evaluation of stability of DMBQ electrode and carbon-only electrode in LiCl/DMSO electrolyte solution during redox. (a) Cyclic voltammogram and (b) Cole-Cole plot of carbon-only electrode without DMBQ. (c) Cyclic voltammogram and (d) Cole-Cole plot of DMBQ electrode.
In contrast, the reduction/oxidation peak current of the DMBQ electrode decreased in accordance with the redox cycle, and the reduction peaks shifted to the negative direction, as shown in Fig. 5c, which accompanied significant increase in the resistance as shown in Fig. 5d. This is consistent with the discharge-charge cycle characteristic trend shown in Fig. 1b, suggesting that DMBQ is gradually degrading and degenerating in the LiCl/DMSO electrolyte solution. Low-molecular-weight active materials are often said to elute into the electrolyte solution, causing cycle degradation,32–37 which is similar to this study.
It has been reported that DMSO-based electrolytes are not stable to lithium metal.38,39 However, according to a previous report using LiCl/DMSO solution, the poly(butyl viologen dichloride)-based electrode showed stable discharge-charge cycles.8 Moreover, Li-Li symmetric cells with LiCl/DMSO solution showed a rather small increase in overvoltage below 20 mV for up to 500 h when cycled at a current density of 0.1 mA cm−2.8 Judging from these results, although the DMSO content slightly decreased in this study (as shown in Table 1), the effect of the change in lithium counter electrode on cycle degradation is considered to be minor.
These findings suggest that the main cause of the cycle degradation in this study was not the degradation of the electrolyte or the lithium counter electrode, but rather the denaturation of DMBQ by the redox cycle and its elution into the LiCl/DMSO electrolyte solution.
We investigated the electrochemical properties and redox reaction mechanism of DMBQ in LiCl/DMSO electrolyte solution. Our main findings are listed below. Not only p-type organic compounds8 but also n-type organic compounds such as DMBQ showed electrochemical reactions in LiCl/DMSO electrolyte solution, suggesting that the electrolyte solution may be applicable to various battery systems.
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.28528721.
Sho Okubo: Conceptualization (Lead), Investigation (Lead), Writing – original draft (Lead)
Masahiko Hayashi: Methodology (Lead)
Hiroaki Taguchi: Project administration (Lead)
Atsushi Aratake: Supervision (Lead)
The authors declare no conflict of interest in the manuscript.
S. Okubo, M. Hayashi, and H. Taguchi: ECSJ Active Members
A part of this paper has been presented in the PRiME2024 (Presentation A07-0838) and the 91st ECSJ Annual Meeting (Presentation S8-3_3_06) in 2024.
