Electronic interaction between dimethyl carbonate and Li⁺ studied by attenuated total reflectance far-ultraviolet spectroscopy

Abstract

Organic electrolytes with Li⁺ were analyzed by far-ultraviolet (≤200 nm) spectroscopy, achieved by an attenuated total reflectance setup. The spectra showed a redshift with Li⁺ addition, attributed to the charge transfer, as revealed by quantum chemical calculations. Multivariate analysis successfully decomposed the spectra into pure solvent and Li-coordinated solvent components.

---

In the development of lithium-ion batteries, the combinations of electrode materials, Li⁺ salts, and electrolytes are important.^1,2 In particular, the electrolyte plays a crucial role as a medium through which Li⁺ diffuses between the cathode and anode.³ Within the electrolyte, understanding the Li⁺ solvation structure is fundamentally significant, and it has been studied using techniques such as mass spectroscopy,⁴ and Raman spectroscopy,^5,6 as well as density functional theory (DFT) and classical molecular dynamics (MD) simulations.^7–9 While the coordination number (CN) in the Li⁺ solvate structures remains a topic of debate, the typical coordination CN is most likely 4 due to the energetic advantage.^8,9

Dimethyl carbonate (DMC), a typical linear carbonate, has attracted attention as an environmentally friendly solvent due to its low volatility and low toxicity.^3,10–12 The intermolecular structure, dynamics, and other properties of DMC have been studied.^13–16 MD simulations show that the CN to Li⁺ in 0.1–1.0 M lithium tetrafluoroborate (Li[BF₄]) DMC solution is approximately 4.¹⁵ Li et al. studied the Li⁺ solvation structures in lithium bis(trifluoromethanesulfonyl)imide (Li[TFSI]) DMC electrolytes using DFT calculation and Raman spectroscopy.¹⁶

In electronic devices, the electronic states of electrolytes are essentially important. Since the 1970s, researchers have investigated the electronic states of DMC using photoelectron spectroscopy and quantum chemical calculation.^17,18 Although ultraviolet (UV) spectroscopy is a powerful method for examining electronic states without ionizing molecules, there have been few papers on this topic. Recently, Das et al. reported the absorption spectra of DMC in the energy range of 7–11 eV using synchrotron radiation, along with the attribution of spectral bands through quantum chemical calculations.¹⁹ However, it's worth noting that in their study, spectral measurements were conducted in a vacuum, placing DMC in the gas phase. This is because water vapor and oxygen in the atmosphere absorb light below 200 nm. Given that DMC is utilized as an electrolyte in the liquid phase, an alternative spectroscopic method is required to capture absorption spectra of liquid samples below 200 nm. Additionally, UV spectroscopic investigations regarding the interaction between DMC and Li⁺ are more challenging.

In 2007, attenuated total reflectance (ATR)-based UV spectroscopy was developed.²⁰ In this technique, light from a D₂ lamp is directed into the sapphire ATR prism, and the evanescent wave generated at the interface between the prism and samples was used as the probe light. In this instrument, the introduction of N₂ gas along the optical path allows for the exclusion of the effects of water vapor and oxygen absorption. It allows the measurement of spectra in the far-ultraviolet (FUV, ≤200 nm) region without the need for a vacuum environment. Thus, the introduction of ATR technology into the FUV region has expanded the range of target compounds and provided information on various electronic structures and transitions.^21,22 Additionally, the practicality of quantum approaches in molecular spectroscopy is steadily increasing, and the ability to analyze spectra is improving.²³ Recently, studies employing ATR-FUV on ionic liquids, which have gained attention as new electrolytes, were reported.^24,25 Additionally, investigations into the electronic interactions between Li⁺ and ionic liquids²⁶ or polymers²⁷ were conducted. In these studies, assignments based on time-dependent density functional theory (TD-DFT) calculations were performed.

In this study, electrolytes comprising DMC and Li⁺ were analyzed using ATR-FUV spectroscopy. The ATR spectra exhibit absorption only in the FUV region, which cannot be observed by conventional spectroscopy. TD-DFT calculations revealed that DMC coordinated to Li⁺ (hereinafter referred to as Li⁺DMC) shows absorbance at the longer wavelength compared to free DMC. Furthermore, multivariate analysis successfully demonstrated that absorbances attributable to free DMC and Li⁺DMC linearly decrease and increase, respectively, with Li⁺ concentration.

DMC, Li[BF₄] and Li[TFSI] were purchased from Kanto Chemical Co., Inc (Tokyo, Japan), and used without further purification. Li[BF₄] and Li[TFSI] solutions of various concentrations (0.0, 0.5, 1.0 and 1.5 M) were prepared. These solutions were placed on a sapphire prism (optical path length = 8 mm, UV grade, Optoline, Tokyo, Japan) and ATR absorption spectra were recorded in the 145–300 nm wavelength range at 0.1 nm intervals (scan rate = 150 nm min⁻¹, incidence angle = 70°). ATR absorption was quantified as −log(I/I₀), where I and I₀ represented the light intensity reflected from the ATR prism in the presence and absence of the sample, respectively.

DFT and TD-DFT calculations were conducted based on models of DMC and Li⁺DMC to assign the measured ATR absorption spectra. Energy minimization of the ground states was performed using the B3LYP method with the cc-pVTZ basis set. Their perpendicular transition energies were subsequently calculated using the TD-CAM (Coulomb-attenuation method)-B3LYP method with the aug-cc-pVTZ basis set. The Gaussian 16 (Version 1.1) program was used for the calculations. Based on the calculated vertical transition energies and oscillator intensities, the absorbance spectra were simulated, assuming an energy width of 0.5 eV for each transition. All molecular orbitals were obtained under the same conditions, with the equivalent number set to 0.02.

Subsequently, the acquired spectra were analyzed via multivariate analysis, with multivariate curve resolution-alternating least squares (MCR-ALS, with non-negative and spectrum constraints), aiming at spectral decomposition into two components – the spectra of pure DMC and Li⁺DMC. The details are described later.

Fig. 1(a) and (b) show the ATR-FUV spectra and their difference spectra of Li[BF₄] in DMC solvent within the 145–300 nm range, respectively. In Fig. 1(a), all samples exhibit a prominent absorption peak at 150 nm, and the absorbance decreases with increasing Li⁺ concentration. The addition of Li⁺ causes the coordination of DMC to Li⁺ (Li⁺DMC). The coordination of DMC with Li⁺ reduces the quantity of uncoordinated free DMC while increasing the quantity of Li⁺DMC. Therefore, the decrease in peak intensity near 150 nm can be attributed to the decrease in free DMC-derived absorption. It should be noted here that the ATR spectrum includes the effect of both the refractive index (n) and the extinction coefficient (κ) of both the sample and prism.²⁰ Consequently, the ATR spectra differ from the transmission spectra, which only consider the effect of κ. In order to separate the effects of n and κ on the ATR spectra, they should be subject to the Kramers–Kronig transform (KKT). However, it is difficult to apply KKT when the absorbance of the sample is strong or when the spectral intensity is not zero at the edge of the measurement wavelength range.²¹ In this study, the acquired ATR-FUV spectra were used without KKT, as discussed later in SI 1 (ESI†).

Fig. 1 (a) ATR-FUV absorption spectra of pure DMC (black line) and Li[BF₄] DMC solutions (colored lines) and (b) difference spectra in the 145–300 nm spectral region. (c) Normalized spectra of Fig. 1(a) and (d) their difference spectra. (e) Li⁺ concentration dependence of the area ratio of the ATR spectral intensity in 153–157 nm to 148–152 nm in Fig. 1(a). The standard deviation bar was calculated from the results of four ATR measurements.

Subsequently, to analyze the Li⁺DMC-derived absorption peak, normalization was performed using the maximum intensity. Fig. 1(c) and (d) represent the normalized spectra of Fig. 1(a) and their difference spectra, respectively. An increase in peak intensity near 155 nm was observed with increasing Li⁺ concentration. As the Li⁺ concentration increases, the amounts of Li⁺DMC increase, suggesting that the absorption at 155 nm is attributed to Li⁺DMC. Therefore, it can be concluded that Li⁺DMC has an absorbance at longer wavelengths (∼155 nm) than free DMC (∼150 nm).

Fig. 1(e) shows the Li⁺ concentration dependence of the area ratio of the ATR spectral intensity in 153–157 nm to 148–152 nm in Fig. 1(a). As discussed above, the former absorbance (153–157 nm) is due to Li⁺DMC, and the latter (148–152 nm) was due to free DMC. In the present concentration range (0–1.5 M), the area ratio was linearly related, indicating the uniform solvation structure. For the 1.5 M Li[BF₄] solution, the molar ratio of DMC to Li⁺ is approximately 8. The reported CN of DMC to Li⁺ is approximately 2.^28,29 Therefore, the number of DMC molecules greatly exceeds the number required for coordinating with Li⁺ at the present Li[BF₄] concentrations in this study. According to the previous papers,^28,29 [BF₄]⁻ also coordinated to Li⁺. The influence of anions is currently being investigated using not only ATR-FUV spectroscopy but also other techniques such as Raman spectroscopy and MD simulations, which are beyond the scope of this paper.

The TD-DFT calculation results supported the experimental findings as follows. Fig. 2(a) displays the calculated oscillator strengths and molar extinction coefficients (ε) of DMC (black line) and Li⁺DMC (red line). Both DMC and Li⁺DMC have absorption in the FUV region, with Li⁺DMC showing absorbance at a longer wavelength compared to DMC, consistent with the experimental results. As discussed in SI 1 (ESI†), there are slight differences in wavelength between the experimental and simulation results. Nevertheless, it should be emphasized that the experimental findings, namely the appearance of absorption in the FUV region and the redshift of the absorption wavelength in the presence of Li⁺, are well reproduced by TD-DFT calculations. The effect of [BF₄]⁻ was considered in SI 2 (ESI†). In short, the effect of [BF₄]⁻ on the ATR spectra in this study is negligibly small. Fig. 2(b) shows the main initial and final molecular orbitals of the primary oscillator strengths. In the case of the DMC model, the main calculated electronic transitions (140.1 and 137.1 nm) are due to the intramolecular excitation of DMC. On the other hand, in the case of Li⁺DMC, the major oscillator strengths (148.1 and 142.8 nm) are mainly due to electron transfer from DMC to Li⁺. Therefore, it can be concluded that the experimentally observed spectral changes in the FUV region with increasing Li⁺ concentration are due to the differences in wavelength between the intramolecular excitation in DMC and the electronic transition from DMC to Li⁺. More details of the TD-DFT results are summarized in SI 3 (ESI†).

Fig. 2 (a) Time-dependent density functional theory-calculated oscillator strengths and molar extinction coefficients (ε) of DMC and Li⁺DMC. (b) Main initial and final molecular orbitals of the electronic transitions.

Next, a multivariate analysis was conducted. It was presumed that the experimentally obtained spectrum (Fig. 1(a)) could be decomposed into two spectra, which is supported by the estimation of the number of components in the singular value decomposition (Fig. S5, ESI†). These two spectra correspond to the spectrum of free DMC (component 1) and that of Li⁺DMC (component 2). Component 3 was added as a noise component, and MCR-ALS was applied. As a constraint during the least-squares calculation, the shape of component 1 was fixed to the experimental spectrum of pure DMC (Fig. 1(a), black line), and the calculations were performed under non-negative value conditions. Further details of the analysis are described in SI 4 (ESI†).

The spectra (S_i^T) and contribution (C_i) of each component (i = 1–3) obtained by MCR-ALS are shown in Fig. 3(a) and (b), respectively. As shown in Fig. 3(a), component 2 has a peak at ∼156 nm, which is a longer wavelength than component 1. As discussed in Fig. 1 and 2, Li⁺DMC exhibits absorbance at a longer wavelength compared to free DMC. The MCR-ALS calculations also indicated the same conclusion.

Fig. 3 (a) and (b) Decomposed absorption spectra (S_i^T) and their contributions vs. Li⁺ concentrations (C_i) analyzed using multivariate curve resolution-alternating least squares (MCR-ALS) calculations for i = 1, 2, and 3. (c) The S_i^TC_i spectra (i = 1–3).

As shown in Fig. 3(b), the contribution of the free DMC-derived absorption (component 1) decreased in proportion to the Li⁺ concentration, while the Li⁺DMC-derived absorption (component 2) increased. In addition, the noise component (component 3) is negligibly small. The S_i^TC_i spectra (i = 1–3) are summarized in Fig. 3(c). The values of Component 3 are at the noise level. These results mean that the experimentally obtained spectra were successfully decomposed into two components (free DMC and Li⁺DMC).

Fig. 4(a) shows the ATR spectra of Li[TFSI] in DMC solvent in the 145–200 nm region. The difference spectra (Fig. 4(b)) indicate a decrease in spectral intensity around 150 nm and an increase around 158 nm with increasing Li⁺ concentration. The normalized spectra (Fig. S6(a), ESI†) and the difference spectra (Fig. S6(b), ESI†) also reveal an increase in the absorption band around 158 nm with the addition of Li⁺. Therefore, in both cases of Li[TFSI] and Li[BF₄], the absorption wavelength of Li⁺DMC is longer than that of free DMC. Additionally, the introduction of Li⁺ to DMC decreases the absorption attributed to free DMC and increases that attributed to Li⁺DMC. It should be noted here that the addition of Li[BF₄] resulted in a decrease in the spectral intensity in the 145–170 nm range (Fig. 1(a) and (b)), while the introduction of Li[TFSI] resulted in a decrease below 150 nm and an increase in the 150–175 nm range (Fig. 4(a) and (b)). These differences are due to the absorbance of [TFSI]⁻ in the FUV region. The effect of [TFSI]⁻ and the multivariate analysis of the Li[TFSI] solution are described in SI 6 and 7 (ESI†), respectively.

Fig. 4 (a) Absorption spectra of pure DMC (black line) and Li[TFSI] DMC solutions (colored lines) at each Li⁺ concentration in the 145–200 nm spectral region obtained via ATR-FUV spectroscopy. (b) Difference spectrum from 0 M at each Li⁺ concentration.

Additionally, the case of Li[PF₆] solution in DMC solvent is summarized in SI 8 (ESI†). [PF₆]⁻ has no absorbance in the 145–180 nm region, and the concentration dependence of Li[PF₆] on the spectra shows similar results with the case of Li[BF₄].

In summary, the analysis revealed the electronic structure of DMC and Li⁺DMC using ATR-FUV spectroscopy, TD-DFT calculations, and MCR-ALS calculations. The ATR spectra showed a decrease in the intensity of the absorption peak of DMC near 150 nm with increasing Li⁺ concentration. TD-DFT calculations indicated that the absorption peak of Li⁺DMC was at a longer wavelength than that of the free DMC. It was also suggested that the redshift was due to the difference in wavelength between the intramolecular excitation of DMC and the electronic transition from DMC to Li⁺. Subsequently, the MCR-ALS algorithm successfully decomposed the obtained ATR spectra into pure DMC and Li⁺DMC. This study represents the first elucidation of the electronic state of DMC in the liquid phase in the FUV region, which cannot be measured by conventional electronic absorption spectroscopy, through ATR-FUV spectroscopy.

This study was financially supported by JSPS KAKENHI, grant number: 23K04811. H. S. mainly prepared the manuscript, and did spectroscopic experiments and quantum chemical calculations. N. U. performed multivariate analysis, and prepared the manuscript for the relevant section. I. T. designed and conducted the study.

Conflicts of interest

There are no conflicts to declare.

Notes and references

1. W. Li, Z. Yang, J. Cheng, X. Zhong, L. Gub and Y. Yu, Nanoscale, 2014, 6, 4532.
2. X. Zhou, J. Xu, D. Li, S. Liu, Y. Chen, C. Zheng and W. Xie, Mater. Sci. Eng., 2018, 394, 42033.
3. K. Xu, Chem. Rev., 2004, 104, 4303.
4. A. von Wald Cresce, O. Borodin and K. Xu, J. Phys. Chem. C, 2012, 116, 26111.
5. Y. Yamada, Y. Koyama, T. Abe and Z. Ogumi, J. Phys. Chem. C, 2009, 113, 8948.
6. M. G. Giorgini, K. Futamatagawa, H. Torii, M. Musso and S. Cerini, J. Phys. Chem. Lett., 2015, 6, 3296.
7. I. Skarmoutsos, V. Ponnuchamy, V. Vetere and S. Mossa, J. Phys. Chem. C, 2015, 119, 4502.
8. O. Borodin, M. Olguin, P. Ganesh, P. R. C. Kent, J. L. Allen and W. A. Henderson, Phys. Chem. Chem. Phys., 2016, 18, 164.
9. W. Cui, Y. Lansac, H. Lee, S. T. Hong and Y. H. Jang, Phys. Chem. Chem. Phys., 2016, 18, 23607.
10. P. Tundo and M. Selva, Acc. Chem. Res., 2002, 35, 706.
11. G. Righi, P. Bovicelli, M. Barontini and I. Tirotta, Green Chem., 2012, 14, 495.
12. Bruno A. V. Santos, Viviana M. T. M. Silva, José M. Loureiro and Alírio E. Rodrigues, ChemBioEng Rev., 2014, 1, 214.
13. S. K. Reddy and S. Balasubramanian, J. Phys. Chem. B, 2012, 116, 14892.
14. X. Wu, X. Zhou, P. Hemberger and A. Bodi, J. Phys. Chem. A, 2017, 121, 2748.
15. J. Soetens, C. Millot and B. Maigret, J. Phys. Chem. A, 1998, 102, 1055.
16. L. Wang, Z. Luo, H. Xu, N. Piao, Z. Chen, G. Tian and X. He, RSC Adv., 2019, 9, 41837.
17. S. P. McGlynn and J. L. Meeks, J. Electron Spectrosc. Relat. Phenom., 1975, 6, 269.
18. S. P. McGlynn and J. L. Meeks, J. Electron Spectrosc. Relat. Phenom., 1976, 8, 85.
19. A. K. Das, S. Krishnakumar and B. N. Rajasekhara, J. Quant. Spectrosc. Radiat. Transfer, 2018, 217, 116.
20. N. Higashi, A. Ikehata and Y. Ozaki, Rev. Sci. Instrum., 2007, 78, 103107.
21. Y. Ozaki, K. B. Beć, Y. Morisawa, S. Yamamoto, I. Tanabe, C. W. Huck and T. S. Hofer, Chem. Soc. Rev., 2021, 50, 10917.
22. Y. Ozaki, Y. Morisawa and I. Tanabe, Chem. Soc. Rev., 2024, 53, 1730.
23. Y. Ozaki, Y. Morisawa, I. Tanabe and K. B. Beć, Spectrochim. Acta, Part A, 2021, 253, 119549.
24. I. Tanabe, A. Suyama, T. Sato and K. Fukui, Analyst, 2018, 143, 2539.
25. I. Tanabe, A. Suyama, T. Sato and K. Fukui, Anal. Chem., 2019, 91, 3436.
26. M. Imai, I. Tanabe, A. Ikehata and Y. Ozaki, Phys. Chem. Chem. Phys., 2020, 22, 21768.
27. N. Ueno, T. Wakabayashi, H. Sato and Y. Morisawa, J. Phys. Chem. A, 2019, 123, 10746.
28. N. Chapman, O. Borodin, T. Yoon, C. C. Nguyen and B. L. Lucht, J. Phys. Chem. C, 2017, 121, 2135.
29. Z. Cao, M. Hashinokuchi, T. Doi and M. Inaba, J. Electrochem. Soc., 2019, 166, A82.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4cc01775g

---