Halide-free deep eutectic solvents with low viscosity and corrosion for efficient SO₂ capture and conversion under environmental conditions

Abstract

Eight halogen-free deep eutectic solvents (DESs) with low viscosity were synthesized for the capture and subsequent catalytic conversion of SO₂. The resulting DESs exhibit excellent SO₂ solubility (up to 1.2514 g g⁻¹ at 20 °C and 1 bar) and high catalytic activity in the cycloaddition between SO₂ and epoxides. Spectroscopic and theoretical investigations elucidate the mechanisms of SO₂ capture and conversion, revealing a synergistic effect between hydrogen bond donors (HBDs) and hydrogen bond acceptors (HBAs).

---

Sulfur dioxide (SO₂), a well-known respiratory irritant and corrosive gas, presents substantial environmental and public health hazards when emitted directly, leading to ecosystem deterioration and serious health issues in humans.^1–3 Conventional flue gas desulfurization (FGD) technologies, encompassing dry scrubbing, wet limestone processes, and biological treatment systems, present significant limitations despite their widespread application. These methods are often accompanied by major challenges such as the formation of secondary pollutants, severe equipment corrosion, and ongoing difficulties in wastewater treatment.^4,5 Therefore, the development of efficient SO₂ capture techniques is essential for environmental protection.⁶

Solvent-based systems show promise for SO₂ capture, achieving high removal efficiencies while ensuring operational stability and exhibiting strong resistance to impurities under complex flue gas conditions.⁷ As emerging green solvents, the DESs have advantages including low cost, high thermal stability, straightforward synthesis without purification, and minimal ecotoxicity.^8–10 Their tuneable properties through strategic selection of hydrogen bond donors (HBDs) and hydrogen bond acceptors (HBAs) enable tailored designs for specific applications, and consequently they have been widely applied in various fields.¹¹ For example, Wang et al.¹² developed a 1-ethyl-3-methylimidazole chloride (EmimCl) and pyridine derivative deep eutectic solvent, where the EmimCl : 2-NH₂Py (7 : 1) system demonstrates rapid absorption rate, high selectivity, and superior recyclability. Despite notable progress in SO₂ capture using DESs, the majority of these systems require energy-intensive high-temperature regeneration processes.^13–15 Direct conversion of absorbed SO₂ into value-added chemicals could simultaneously mitigate regeneration energy demands and simplify downstream processing.¹⁶ Our prior work¹⁷ established an EmimCl-Trtz (2 : 1) DES system that concurrently absorbed SO₂ and catalysed its cycloaddition with epoxides into cyclic sulfites, effectively addressing both absorbent regeneration and value-added product synthesis. Nevertheless, the high viscosity and toxic, corrosive halogen components of the DESs hinder SO₂ absorption efficiency and compromise green sustainable conversion processes.^18,19 Hence, the rational design of low-viscosity, halogen-free dual-functional DESs that synergistically combine efficient SO₂ capture with catalytic epoxide conversion is imperative to advance sustainable and clean chemical processes.

Inspired by the polyhydroxycarboxylic acid ionic liquid-catalysed CO₂ cycloaddition reaction,²⁰ we developed halogen-free DESs for efficient SO₂ capture and subsequent conversion. As illustrated in Fig. 1, eight DESs were prepared using 1-ethyl-3-methylimidazolium acetate (EmimOAc) and N-alkylimidazolium acetate proton ionic liquids (PILs) as HBAs, paired with 2-aminopyridine (2-AmPy) and imidazole (Im) as HBDs, at a 1 : 1 molar ratio. The nuclear magnetic resonance hydrogen spectrum (¹H NMR) indicates that ethyl imidazole and acetic acid have undergone a complete neutralization reaction, forming EimOAc as the HBA (Fig. S1). Thereafter, the physicochemical properties of these DESs were evaluated using thermogravimetric analysis (TGA, Fig. S2), viscometry, and densitometry, while structural characteristics were analysed by NMR (Fig. S3–S10) and Fourier transform infrared (FTIR, Fig. S11) spectroscopy. Among these DESs, [EmimOAc][lm] demonstrates the best thermal stability, with an initial decomposition temperature of 120.8 °C. Notably, the viscosities of the six PIL-based DESs are all below 20 cP (Table S1), with [EimOAc][Im] exhibiting the lowest viscosity of 11.8 cP at 20 °C, which facilitates SO₂ absorption.

Fig. 1 Chemical structures of HBAs and HBDs for DESs.

Initially, we evaluated the absorption capacity of eight DESs at 20 °C by introducing SO₂ into 1.0 g of DES. As depicted in Fig. 2a, all DESs rapidly absorbed SO₂ within 5 min and reached absorption equilibrium in approximately 10 min. The absorption capacities of these DESs all exceeded 0.9 g g⁻¹, with [EmimOAc][Im] demonstrating a superior absorption capacity of 1.2514 g g⁻¹, which is comparable to or surpasses those of previously reported DESs (Table S2). The influence of HBAs on absorption performance was further investigated. Fig. 2b demonstrates that the SO₂ solubility follows this descending order: [EmimOAc][Im] > [MimOAc][Im] > [EimOAc][Im] > [BimOAc][Im], demonstrating that non-protonic ionic liquid exhibits significantly higher SO₂ absorption capacity than PILs. Furthermore, a shorter alkyl chain in the cation of PILs is beneficial for SO₂ absorption. Im demonstrates stronger alkalinity than 2-AmPy, thereby enhancing SO₂ uptake performance in all HBD-based DESs except [BimOAc][Im]. Nevertheless, the uptake variations among these DESs remain relatively minor.

Fig. 2 (a) and (b) Comparison of SO₂ absorption in eight DESs at 20 °C and 1.0 bar. (c) the effects of temperature at 1.0 bar, and (d) SO₂ partial pressure at 30 °C on SO₂ absorption.

The influence of temperature and pressure on SO₂ absorption by the three DESs was investigated. The results shown in Fig. 2c and d reveal that the solubility of SO₂ in all three DESs decreases as the temperature increases and pressure decreases. For instance, the SO₂ absorption capacity of [EmimOAc][Im] decreased from 1.2514 g g⁻¹ at 20 °C to 0.7588 g g⁻¹ at 60 °C under 1.0 bar. This decline can be attributed to the fact that an increase in temperature not only impedes the physical solubilization of SO₂, as described by Henry's law, but also weakens the acid–base interactions between the DESs and SO₂, ultimately resulting in reduced solubility of SO₂. Notably, SO₂ absorption shows a distinct nonlinear pattern at low pressures (0–0.2 bar), typically attributed to chemical absorption. Beyond 0.2 bar, absorption increases nearly linearly, indicating physical absorption as the dominant process. At a partial pressure of 0.2 bar SO₂, [EmimOAc][Im] exhibits an absorption capacity of 0.5812 g g⁻¹, demonstrating significant SO₂ uptake even at lower pressures. To further investigate the absorption capacity of [EimOAc][Im] at low SO₂ concentrations, we evaluated their performance at 6000 ppm and 1000 ppm, achieving absorption capacities of 0.336 and 0.073 g g⁻¹ (Fig. S12). The regeneration performance of DESs is crucial in the process of SO₂ capture. We thus conducted absorption-desorption cycling experiments at temperatures of 30 °C and 80 °C to evaluate the DESs. The results shown in Fig. S13 demonstrate that [EimOAc][Im] retains its efficient SO₂ capture capacity after five cycles.

Corrosion tests (weight loss method) using Q235 iron spheres show that the halogen-free [EmimOAc][2-AmPy] causes significantly less corrosion than [EmimBr][2-AmPy] (Fig. 3 and Table S3). Q235 iron spheres in [EmimBr][2-AmPy] exhibited substantial mass loss (corrosion rate: 0.2455 g a⁻¹) and visible surface changes, while those in [EmimOAc][2-AmPy] remained largely unaffected (corrosion rate: 0.0011 g a⁻¹). These results demonstrate that halogen-free DESs reduce equipment corrosion, making them more suitable for practical applications.

Fig. 3 Comparison of the mass lost by the 3 iron balls per year. Experimental conditions: DES (2 mL), H₂O (0.1 mL), 3 iron balls (Q235). Mass loss is determined by an analytical balance with an accuracy of ± 0.0001 g.

After evaluating the SO₂ absorption, DESs were used as catalysts for the cycloaddition of SO₂ and styrene oxide (SO) (Table 1). These DESs demonstrated exceptional catalytic performance, achieving yields exceeding 90% for eight DESs at 60 °C and 1 bar SO₂ (entries 1–8). The synergistic catalytic effect of HBA and HBD in the DESs results in enhanced catalytic activity relative to either HBA or HBD alone, as well as their corresponding precursor compounds (Table S4). To systematically optimize the reaction parameters, [MimOAc][2-AmPy], which demonstrated excellent catalytic activity, was selected as the model catalyst. Time-dependent evaluation at 30 °C revealed incremental yield improvements from 84% (8 h) to 85% (12 h), reaching 92% at the optimal duration of 24 h (entries 9–11). Consequently, the optimal conditions were established as 30 °C and 24 h, under which three DESs with 2-AmPy as the hydrogen bond donor (HBD) achieved yields ranging from 84% to 91%, respectively (entries 12-14). Significantly, [EmimOAc][2-AmPy] exhibited the lowest catalytic activity among the three DESs, with alkyl chain elongation demonstrating an inverse correlation with catalytic efficiency. Given the limited thermal stability of [MimOAc][Im], [EimOAc][Im] and [BimOAc][Im] were evaluated (entries 15 and 16), attaining 95% and 93% yields, which may be attributed to Im having a stronger interaction with SO₂ than that of 2-AmPy. The results indicate that [EimOAc][Im] has the best catalytic effect. To evaluate the effect of water content on catalytic performance, we tested [EimOAc][Im] containing 10wt% H₂O (entry 17). The results indicate a minimal impact on catalytic efficiency.

Table 1 Optimization of the reaction conditions for the cycloaddition of styrene oxide (SO) and SO₂^a

Entry	Catalyst	T (°C)	t (h)	Yield^b (%)	Sel.^b (%)
a Reaction conditions: SO (3 mmol), catalyst (0.3 mmol), SO2 (1.0 bar). b Determined by gas chromatography. c 10 wt% H2O.
1	[MimOAc][2-AmPy]	60	5	95	>99
2	[EimOAc][2-AmPy]	60	5	95	>99
3	[BimOAc][2-AmPy]	60	5	95	>99
4	[EmimOAc][2-AmPy]	60	5	92	>99
5	[MimOAc][Im]	60	5	95	>99
6	[EimOAc][Im]	60	5	95	>99
7	[BimOAc][Im]	60	5	94	>99
8	[EmimOAc][Im]	60	5	93	>99
9	[MimOAc][2-AmPy]	30	8	84	96
10	[MimOAc][2-AmPy]	30	12	85	97
11	[MimOAc][2-AmPy]	30	24	92	96
12	[EimOAc][2-AmPy]	30	24	91	95
13	[BimOAc][2-AmPy]	30	24	85	94
14	[EmimOAc][2-AmPy]	30	24	84	94
15	[EimOAc][Im]	30	24	95	>99
16	[BimOAc][Im]	30	24	93	98
17^c	[EimOAc][Im]	30	24	92	>99

---

Under the optimized conditions, the [EimOAc][Im]-catalyzed cycloaddition of diverse epoxides with SO₂ achieved yields spanning 54% to 98% (2a–2h), demonstrating broad substrate compatibility. Fig. 4a shows the [EimOAc][Im]-catalyzed cycloadditions, where epichlorohydrin and 1,2-epoxyoctane with SO₂ delivered yields of 89% and 98% (2a and 2b). Notably, even epoxides with bulky substituents on the carbon atoms, including allyl glycidyl ether, styrene oxide, or benzyl glycidyl ether, maintained exceptional efficiency (≥90% yields, 2d, 2e, 2g). Furthermore, cyclohexene oxides exhibited reduced reactivity due to their structural constraints as internal epoxides compounded by significant steric hindrance at the oxirane ring, ultimately resulting in a diminished yield of 54% (2h). The results indicated that the [EimOAc][Im] catalyst displays excellent catalytic activity.

Fig. 4 (a) The cycloaddition of SO₂ with various epoxides. (b) SO₂ capture and fixation for the synthesis of 2e in [EimOAc][Im].

Upon completion of the reaction, water and ethyl acetate were added to the reaction solution. The [EimOAc][Im] catalyst could then be isolated by removing water from the aqueous phase, allowing for the recovered catalyst to be utilized directly in the subsequent cycle of the reaction. As shown in Fig. S14, the activity of the catalyst decreased by almost 10% in yield after each cycle. The reduced catalytic performance is attributed to partial dissolution of [EimOAc][Im] in ethyl acetate. To improve integration between absorption and conversion processes, we investigated the in situ transformation of absorbed SO₂ into cyclic sulfite. At 60 °C, [EimOAc][Im] achieved SO₂ absorption equilibrium with a capacity of 5.8 mmol, followed by the addition of an equimolar amount of SO (1e). Upon completion of the 5 h reaction, cyclic sulfite 2e was isolated in 95% yield (Fig. 4b). After separation from the product, [EimOAc][Im] can be reused in the SO₂ capture and conversion cycle. The cycling performance of DES was evaluated in situ (Fig. S15). The results show that the conversion efficiency remained above 90% over three cycles. Notably, however, the viscosity of the recovered [EimOAc][Im] increased markedly after the third cycle, likely due to accumulated unreacted cyclic sulfites. This change may impair the mass transfer properties of [EimOAc][Im] and consequently reduce its absorption performance. The above results provide new insights into the coupling of SO₂ capture and conversion processes.

FTIR and NMR spectra of [EimOAc][Im] before and after the absorption of SO₂ were analyzed to investigate the absorption mechanism. As can be seen in Fig. S16, the chemical shifts of the three H atoms of imidazole in [EimOAc][Im] changed from 7.04 (h and i) and 7.70 (g) ppm to 7.52 and 8.75 ppm after absorption of SO₂, indicating that there is an acid–base interaction between SO₂ and Im.²¹ The protons (c–f) on [EimOAc] exhibit downfield shifts after absorption of SO₂, suggesting that [EimOAc] interacts with SO₂ as well. The difference is that the chemical shift of H(1) in [EimOAc][Im] shifted from 10.49 ppm to 9.93 ppm, which is attributed to the absorption of SO₂ affecting the original hydrogen bonding interactions of the DES molecules.²²Fig. 5a displays the FTIR spectra of [EimOAc][Im] before and after SO₂ absorption, revealing four new peaks at 1371, 1179, 968, and 507 cm⁻¹. Among them, the peaks at 1371, 1179 and 507 cm⁻¹ correspond to asymmetrical telescopic vibration, symmetrical telescopic vibration, and bending vibration of S=O, respectively.^23,24 The bond at 968 cm^-1 originates from N–S coordination between the imidazole (Im) group and SO₂.²⁵ Additionally, the carboxyl group peak shifts from 1702 cm⁻¹ to 1723 cm⁻¹ due to the electrophilic environment induced by SO₂, indicating an interaction between the oxygen atoms in the acetate ion's –COO⁻ group and SO₂.²⁶ These spectral signatures confirm a dual trapping mechanism involving physical and chemical absorption in the DES system.

Fig. 5 (a) FTIR spectra of [EimOAc][Im] before and after capturing SO₂. (b) ¹H NMR spectrum of [EimOAc][Im], SO, and [EimOAc][Im] + SO with an internal reference in CDCl₃. (c) Calculated interactions of HBA, HBD, and N sites in [EimOAc][Im] with SO and SO₂. (d) Possible reaction mechanisms for cycloaddition of SO₂ and epoxides.

In order to clarify the catalytic mechanism, we investigated the interaction between epoxide and [EimOAc][Im] using ¹H NMR spectroscopy. As shown in Fig. 5b, when [EimOAc][Im] and styrene oxide (SO) were mixed in equimolar ratios, ¹H NMR analysis revealed a marked upfield shift of the hydrogen-bonded proton H(1) in the DES from 12.18 to 11.23 ppm; however, the chemical shifts of hydrogen atoms in the SO epoxide group remained largely unchanged, suggesting weak hydrogen-bonding interactions between the DES and SO. This differential behavior implies that [EimOAc][Im] likely activates the epoxide through selective hydrogen-bonding interactions with the oxygen atom while minimal electronic perturbation occurs in the SO.

Theoretical calculations were further performed to gain insight into the activation of epoxides and SO₂ by DESs (Fig. 5c). In [EimOAc][Im], after the N atoms on the imidazole ring interacted with SO₂, the bond angle of SO₂ was bent from 113.97° to 113.74°, indicating that SO₂ was effectively activated by the N sites. When SO interacted with proton hydrogens in DES and N-H on the imidazole ring to form hydrogen bonds, the C–O bond length of SO increased from 1.495 Å to 1.527 Å and 1.490 Å to 1.553 Å, respectively. The C–O bond in SO was activated and lengthened accordingly, which was favourable for the ring-opening step during OAc⁻ anion attack. Theoretical calculations demonstrate that the DES facilitates epoxide ring-opening through hydrogen bonding while simultaneously activating SO₂ molecules.

According to both experimental and computational findings, a plausible reaction mechanism is proposed (Fig. 5d). The mechanism initiates with hydrogen bond-mediated activation of the epoxide by DES, inducing polarization of the C–O bond (Int 1). This activation facilitates a regioselective nucleophilic attack by the acetate anion at the electrophilic carbon centre, leading to the formation of an oxyanion intermediate (Int 2). Subsequent insertion of SO₂ into the activated intermediate produces a sulfur-containing adduct (Int 3). This is followed by an intramolecular cyclization step that releases the acetate anion, yielding the target product and allowing for catalyst regeneration.

In summary, we have developed a series of low-viscosity, halogen-free DESs for the simultaneous capture and conversion of SO₂ into cyclic sulfites. These DESs effectively capture SO₂ through Lewis acid–base interactions and charge transfer with acetate, reversibly capturing up to 1.2514 g g⁻¹ (1 bar SO₂ at 20 °C). Remarkably, [EimOAc][Im] demonstrates excellent activity as a catalyst for cycloaddition reactions with various epoxides, yielding moderate to high conversions (54–98%) under mild reaction conditions. Spectroscopic characterization and theoretical calculations revealed a multi-site absorption mechanism and synergistic catalytic process. Our findings not only provide a green and efficient pathway for SO₂ management but also offer insights into the rational design of functional solvents for sustainable chemical processes.

This work was supported by the National Natural Science Foundation of China (no. 22208070 and 22168012).

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information: experimental section and characterization information of DESs. See DOI: https://doi.org/10.1039/d5cc05880e.

Notes and references

1. P. Zhang, G. Xu, M. Shi, Z. Wang, Z. Tu, X. Hu, X. Zhang and Y. Wu, Sep. Purif. Technol., 2022, 286, 120489.
2. J. L. Obeso, C. V. Flores, R. A. Peralta, M. Viniegra, N. Martín-Guaregua, M. T. Huxley, D. Solis-Ibarra, I. A. Ibarra and C. Janiak, Chem. Soc. Rev., 2025, e00997.
3. H. Wang, P. Wu, C. Li, J. Zhang and R. Deng, ACS Sustainable Chem. Eng., 2022, 10, 4451–4461.
4. J. Yan, J. Zhu, S. Tong, Q. Wang and Z. Wang, ACS Appl. Mater. Interfaces, 2024, 16, 56181–56190.
5. 5 F. Liu, J. Yu, A. B. Qazi, L. Zhang and X. Liu, Environ. Sci. Technol., 2021, 55, 1419–1435.
6. R. Zhang, D. Hu, Y. Fu, Q. Feng, C. Mu, K. Gao, H. Ma, M. Liu and M. Zhang, Aggregate, 2024, 5, e408.
7. S. Foorginezhad, G. Yu and X. Ji, Front. Chem., 2022, 10, 951951.
8. H. Zhu, Y. Zhong, L. Yan, H. Zhang, Y. Shen, Z. Le, Q. Fan and Z. Xie, Green Chem., 2024, 26, 1387–1392.
9. Q. Zhang, K. De Oliveira Vigier, S. Royer and F. Jérôme, Chem. Soc. Rev., 2012, 41, 7108–7146.
10. W. Xiong, Y. Lu, C. Li, J. Geng, Y. Wu and X. Hu, Green Chem., 2023, 25, 1898–1907.
11. H. Qin, X. Hu, J. Wang, H. Cheng, L. Chen and Z. Qi, Green Energy Environ., 2020, 5, 8–21.
12. C. Wang, H. Wu, J. Li, J. Zhang, J. Zhang, J. Ding, H. Li, H. Li and W. Zhu, Chem. Eng. J., 2023, 471, 144394.
13. H. Wu, W. Xiong, S. Wen, X. Zhang and S. Zhang, Chem. Commun., 2022, 58, 7801–7804.
14. Z.-M. Li, W.-Q. Gong, J.-F. Li, S.-X. Zhu, D.-J. Tao and Y. Zhou, J. Mol. Liq., 2022, 367, 120521.
15. Y. Zhou, X.-Y. Sang, H. Guan, X.-J. Shu, Z.-H. Xu, M.-S. Sun and M.-Z. Cai, J. Mol. Liq., 2023, 389, 122921.
16. P. Li, X. Wang, T. Zhao, C. Yang, X. Wang and F. Liu, Chem. Eng. J., 2021, 422, 130033.
17. G. Long, C. Yang, X. Yang, T. Zhao, F. Liu and J. Cao, ACS Sustainable Chem. Eng., 2020, 8, 2608–2613.
18. A. W. Kleij, Curr. Opin. Green Sustainability, 2020, 24, 72–81.
19. J. Li, C. Wang, H. Wu, J. Zhang, M. Zhang, Q. Zhu, Y. Fan, H. Liu and W. Zhu, Energy Fuels, 2023, 37, 8051–8056.
20. Y. Jiang, D. Wang, B. Guo, J. Zhao, Z. Zhou, L. Jin and Y. Lei, Mol. Catal., 2023, 551, 113664.
21. X. Li, L. Meng, F. Yang, Z. Yang, J. Li, Y. Chen and X. Ji, Chem. Commun., 2024, 60, 10560–10563.
22. X. Yang, Y. Zhang, F. Liu, P. Chen, T. Zhao and Y. Wu, Sep. Purif. Technol., 2020, 250, 117273.
23. N. Scaglione, L. Wylie, A. Padua and M. Costa Gomes, ACS Sustainable Chem. Eng., 2024, 12, 10486–10497.
24. P. Liu, K. Cai, X. Zhang and T. Zhao, AIChE J., 2022, 68, 17596.
25. C. Wang, G. Cui, X. Luo, Y. Xu, H. Li and S. Dai, J. Am. Chem. Soc., 2011, 133, 11916–11919.
26. J. Zhao, S. Ren, Y. Hou, K. Zhang and W. Wu, Ind. Eng. Chem. Res., 2016, 55, 12919–12928.

---