Peroxide detected in imidazolium-based ionic liquids and approaches for reducing its presence in aqueous and non-aqueous environments

Abstract

Imidazolium-based ionic liquids were discovered to contain micromolar (μM) quantities of a peroxide species. A general approach using catalase (aqueous solution) or a salen–manganese complex (neat IL) for reducing the presence of the peroxide species is described herein.

---

For nearly 20 years, ionic liquids (ILs) have been used in organic chemistry, bioelectrochemistry and biocatalysis as a replacement for traditional solvents due to their expected low toxicity and thermal stability¹ as well as their ability to stabilize protein structures.² These salt-like solvents can be classified as water miscible or water immiscible depending on their cation/anion composition.³ Their unique physical properties, including high ionic conductivity, make them suitable for catalytic and oxidation reactions.⁴ In 2000, Song and co-workers described a Jacobsen catalyst complex used in IL to catalyse an asymmetric epoxidation reaction⁵ and since then this medium has been used in the oxidation of alcohols, olefins, sulphides, halides, and thiols, among others.⁶ Likewise, oxidoreduction reactions in ILs catalysed by oxidases and dehydrogenases have become more prevalent in recent years due to a growing interest in IL-based biofuel cells and biosensors.⁷ In these systems, the solubility and transport of the redox active species are ensured by the IL, creating a microenvironment capable of mimicking the electron-transfer mechanism in biological systems.^4,8 However, certain enzyme substrates, such as NAD(H) used by dehydrogenases, are only slightly soluble in organic solvents and apolar ionic liquids, thereby limiting their use to water miscible ionic liquids.⁹ Among these water miscible ILs, those with dialkyl substituted imidazolium cations (e.g. [C_nMIm]) are used most prominently.¹⁰

The preparation of these ILs, whether commercially or academically, plays a crucial role in the quality of the product. Unlike a traditional organic solvent, it is not possible to purify an ionic liquid using distillation due to its non-volatile nature and any impurities can alter its properties significantly, including the salt's color.¹¹ The factors at play resulting in a colorless product instead of one that is colored are not entirely understood but most likely correspond to impurities already present in the reagents, thermal degradation of the reagents and/or oxidation products.¹¹ For these reasons, our team routinely synthesizes a water-miscible IL, 1,3-dimethylimidazolium dimethyl phosphate, [MMIm][Me₂PO₄], using the approach by Kuhlmann¹² in order to obtain a more optically pure product in contrast to commercial options.¹³ Synthesizing the liquid at low temperatures (≤80 °C), protected from light and under inert atmosphere minimizes the possibility of contamination, especially from water, which can act as a nucleophile during the quaternarization reaction.

For the current study, we discovered the presence of a peroxide species in an aqueous solution of [MMIm][Me₂PO₄] using horseradish peroxidase (HRP) and its substrate, 2,2′-Azino-bis-(3-ethylbenzthiazoline-6-sulfonic acid) (ABTS), as the system of detection.¹⁴ The H₂O₂ was indirectly detected spectrophotometrically at 420 nm, which corresponds to the subsequent oxidation of ABTS by HRP in the presence of H₂O₂. The reaction takes place as follows:

To our knowledge, the formation or existence of peroxide species in ionic liquids has not been previously described in the literature. However, a peroxide species present in an IL should be considered as an important contaminant since even trace amounts of impurities could have a negative effect on chemical reactions. For instance, Heimer and co-workers recently demonstrated an oxidation reaction in different ionic liquids which involves the oxidative folding of peptides containing cysteine-rich regions.¹⁵ Depending on its concentration, the peroxide species present in an IL could interfere with the oxidation mechanism by disrupting disulfide bond formation within the peptide, keeping it locked in an unfolded state or leading to its misfolding.

The original concentration of the peroxide species detected in the sample was calculated by using the method of standard addition for a range of H₂O₂ concentrations from 0–45 μM to take into account the fact that the HRP's activity might be altered in different types of ILs (Fig. 1).

Fig. 1 Dosage of H₂O₂ using standard curves to calculate peroxide concentration in trimethyl phosphate (▲), redistilled MIm (■) and non-redistilled MIm (●). All measures are performed in triplet in 15% (v/v) MIm, [HRP] 1.30 U mL⁻¹, [ABTS] 2 mM, [H₂O₂] 0–45 μM, phosphate buffer pH 7.6, 20 mM at 25 °C.

In order to determine whether the peroxide present in the samples could be a byproduct of IL synthesis, the starting materials used to obtain [MMIm][Me₂PO₄], methylimidazolium [MIm] and trimethyl phosphate [Me₃PO₄], were also tested for peroxide presence (Fig. 1). The non-redistilled MIm was compared to a redistilled MIm. While the trimethyl phosphate showed no significant peroxide species contamination (<0.2 μM), the concentration detected in the redistilled MIm was 63 μM. This result was significantly less than the 767 μM peroxide species concentration found in the non-redistilled MIm. Since it is thought that the color of an ionic liquid is influenced by contaminant compounds,¹⁶ it is important to note that the non-redistilled MIm, containing the highest concentration of the peroxide species, was an intense yellow-colored solution, whereas the redistilled MIm, containing the least amount of peroxide species, was a colorless solution.

Taking into account these results, fresh batches of [MMIm][Me₂PO₄] were synthesized using the redistilled MIm and the non-redistilled MIm. The [MMIm][Me₂PO₄] synthesized with the redistilled MIm had the least amount of peroxide detected, 123 μM, and was a colorless solution similar to the starting material (Fig. 2). The [MMIm][Me₂PO₄] synthesized with the non-redistilled MIm contained as much as 394 μM of peroxide species and was an intensely yellow-colored liquid like the non-redistilled MIm. Indeed, after vacuum evaporation of excess starting materials, the amount of peroxide in each sample had been significantly reduced: 23% for the non-redistilled [MMIm][Me₂PO₄] and 32% for redistilled [MMIm][Me₂PO₄] (Fig. 2).

Fig. 2 (A) H₂O₂ concentration, before (blue bars) and after (yellow bars) vacuum evaporation, for non-redistilled [MMIm][Me₂PO₄] and redistilled [MMIm][Me₂PO₄]. H₂O₂ concentration after addition of catalase is also shown. (B) UV-vis spectra of academic-synthesized [MMIm][Me₃PO₄] treated with salen–manganese complex after 5 min, 120 minutes and 48 hours. All measurements are performed in triplet.

Recent advances in the field have led to the development of distillable protic ILs based on Bronsted acids and base pairs whereby the neutral precursors are formed after proton displacement during distillation and eventually lead to peroxide elimination.^17,18 Nevertheless, imidazolium-type ILs cannot be purified in this way since they retain their charge (aprotic ILs). However, the starting materials used in their fabrication are classically volatile compounds that can be distilled.¹⁹ This results in decomposition of the peroxide species.²⁰ The reduction of excess MIm and Me₃PO₄ after vacuum evaporation was also verified by ¹H-NMR and ³¹P-NMR (ESI†).

Addition of catalase to a [MMIm][Me₂PO₄] sample provided further confirmation that the contaminant in the different ILs was indeed a peroxide species (Fig. 2A). Metalloproteins such as catalase catalyze dismutation reactions by breaking down H₂O₂ into water and oxygen as shown by the following reaction:

Catalase: 2H₂O₂ → 2H₂O + O₂

After catalase treatment, the concentration of H₂O₂ in the aqueous sample of [MMIm][Me₂PO₄] (15% v/v) was found to be negligible as expected. Therefore addition of catalase in aqueous reactions involving ILs should be recommended. For non-aqueous medium applications, salen–manganese complexes can be used to reduce the presence of peroxide species as they exhibit both superoxide dismutase (SOD) and catalase-like activities.¹⁹ As a proof of concept, we used the salen–manganese complex despite the fact that it does not possess the best catalase-like activity.²¹ After 2 hours of salen treatment, a decrease in peroxide species concentration was observed and, after 48 hours, its presence was reduced by 99% from 394 μM to 3 μM (Fig. 2B). While removing peroxide species from the IL was faster using the enzymatic approach, the salen–manganese complex proved to be an efficient method for anhydrous chemical purposes. Furthermore, the oxidation of imidazolium-based ILs by H₂O₂, or ion pairing by nucleophilic attack of O₂⁻˙ with the imidazolium cation, can both give rise to imidazolones and other byproducts.^22,23 Unfortunately, we were unable to identify such a compound. We then compared the concentration of H₂O₂ detected in an older sample, 244 μM, to a more recently synthesized batch of [MMIm][Me₂PO₄]. The presence of peroxide in the latter was 198 μM. This corresponds to a 19% increase in the quantity of the peroxide species in the older batch and also shows that its presence in the ionic liquid increases overtime, probably due to moisture or oxygen dissolution.

In addition, several commercially-available water soluble ILs were tested and their peroxide content measured (Table 1). The data confirms that a peroxide species was detected in each of the samples analysed. While the dicyanamide and acetate-based ILs had lower quantities of peroxide, the phosphate and sulphate-based ILs showed varying results depending on the sample. The age of the IL is also important as the frequency with which the IL is exposed to light and moisture seems to increase the amount of peroxide present in the sample over time.

Table 1 Detection of peroxide concentration in a selection of commercially available water-miscible ionic liquids and their determined peroxide concentrations^a

Type of IL	[H2O2] μM
a Reaction conditions: 15% (v/v) IL, [HRP] 1.30 U mL^−1, [ABTS] 2 mM, [H2O2] 0–45 μM, phosphate buffer pH 7.6, 20 mM, 25 °C.
1,3-Dimethylimidazolium dimethyl phosphate	34 ± 0.04
1-Ethyl-3-methylimidazolium diethyl phosphate	155 ± 0.04
1-Butyl-3-methylimidazolium hexafluorophosphate	12 ± 0.04
1-Ethyl-3-methylimidazolium ethyl sulfate	57 ± 0.04
1-Butyl-3-methylimidazolium octyl sulfate	177 ± 0.04
1-Butyl-3-methylimidazolium acetate	7 ± 0.04
1-Butyl-3-methylimidazolium dicyanamide	22 ± 0.04
1-Butyl-1-methylpyrrolidinium dicyanamide	27 ± 0.04

---

We propose a mechanism in which the natural composition of the IL is sufficient enough to promote peroxide species formation (Scheme 1). It is well-established that the 2-position of imidazolium rings can readily deprotonate to form a stabilized carbon (II).²⁴ We propose this carbon could react with dissolved molecular oxygen to form an intermediate (III). This may occur directly or via the formation of an endoperoxide as it was described for the oxidation of histidine by singlet molecular oxygen.²⁵ The resulting intermediate could be further stabilized to the peroxide species (IV) and accumulate over time in the IL as oxygen continues to dissolve in the absence of water. This peroxide species could degrade into an imidazolone or hydroxy-imidazolone species in the presence of water.²⁵ Such degradation products could appear after long-time storage and could participate in the apparition of color in the ILs. Their concentration (a few hundred μM) compared to the concentration of the neat IL (some molar) make them highly difficult to isolate. In addition, this could explain why Gu and co-workers observed the formation of hydrogen peroxide during the alkylation reaction of a cyclohexanone with an indole using a Brønsted acid ionic liquid.²⁶ This reaction works best in the presence of oxygen, which could become activated by the IL, thereby forming the peroxide species. In this way, the peroxide species could have originated directly from the IL, instead of resulting from the reaction itself.

Scheme 1 Proposed mechanism for generating a peroxide species from imidazolium-type IL.

Organic reactions using in situ generation of hydrogen peroxide in ILs have been demonstrated and are proposed as an alternative to the industrial use of toxic, non-specific oxidants.²⁷ While ILs have already been used in the polymerization of arenes²⁸ and the reduction of dimethyl maleate and benzaldehyde,²⁹ water-containing ILs have recently been shown to generate hydrogen peroxide by electroreduction of oxygen.³⁰ The in situ generated hydrogen peroxide can then be used for the epoxidation of alkenes for example.³⁰ Oxygen activated by imidazolium-based ILs could offer a catalytic-free alternative for these reactions.

In conclusion, we have shown for the first time the presence of a peroxide species in imidazolium-based ILs used in organic synthesis, biocatalysis and (bio)electronics. In light of these results, some previous studies dealing with oxidation reactions (electrochemical, chemical and biochemical) should possibly be regarded from a different perspective. Peroxides are normally unstable at room temperature but it is possible that room-temperature ILs might stabilize peroxide structures, thereby contributing to peroxide species propagation over time. Furthermore, the peroxide species present in starting materials greatly contributes to the quantity of peroxide species present in the IL. Distillation and vacuum evaporation can help to reduce the amount of peroxide found in starting materials and, consequently, in the finished products. We have also demonstrated that addition of catalase or salen–manganese complexes can eliminate the peroxide species in the IL. While the mechanism of peroxide formation in this medium remains unknown, these findings could help to improve reaction performance in ILs and to reduce unwanted side-reactions.

Acknowledgements

The French Ministry of Education, Research and Technology is gratefully acknowledged for the fellowship of J.L.L.C. Marie-Christine Duclos is kindly acknowledged for providing the salen complex. This work was financially supported by the GDR Synthèse et Procédés durables pour une Chimie éco-compatible.

Notes and references

1. F. Van Rantwijk and R. Sheldon, Chem. Rev., 2007, 107, 2757–2785.
2. K. Fujita, M. Forsyth, D. R. MacFarlane, R. W. Reid and G. D. Elliot, Biotechnol. Bioeng., 2006, 94, 1209–1213.
3. R. A. Sheldon, R. M. Lau, M. Sorgedrager, F. van Rantwijk and K. Seddon, Green Chem., 2002, 4, 147–151.
4. M. Bidikoudi, L. Zubeir and P. Falaras, J. Mater. Chem. A, 2014, 2, 15326–15336.
5. C. E. Song, E. J. Roh, B. M. Yu, D. Y. Chi, S. C. Kim and K. J. Lee, Chem. Commun., 2000, 615–616.
6. D. Betz, P. Altmann, M. Cokoja, W. Herrmann and F. Kühn, Coord. Chem. Rev., 2011, 255, 1518–1540.
7. P. Pinto, M. Saraiva and J. Lima, Anal. Methods, 2008, 24, 1231–1238.
8. M. J. A. Shiddiky and A. A. J. Torriero, Biosens. Bioelectron., 2011, 26, 1775–1787.
9. M. Bekhouche, B. Doumèche and L. J. Blum, J. Mol. Catal. B: Enzym., 2010, 65, 73–78.
10. V. Emel'yanenko, S. Verevkin, A. Heintz, K. Voss and A. Shultz, J. Phys. Chem. B, 2009, 113, 9871–9876.
11. P. Wasserscheid and T. Welton, Ionic Liquids in Synthesis, Wiley-VCH, New York, 2002.
12. E. Kuhlmann, S. Himmler, H. Gielbelhaus and P. Wasserscheid, Green Chem., 2006, 9, 233–242.
13. J. L. L. Carter, M. Bekhouche, A. Noiriel, L. J. Blum and B. Doumèche, ChemBioChem, 2014, 15, 2710–2718.
14. H. Gallati, J. Clin. Chem. Clin. Biochem., 1979, 17, 1–7.
15. P. Heimer, A. Tietze, M. Böhm, R. Giernoth, A. Kuchenbuch, A. Stark, E. Leipold, S. Heinemann, C. Kandt and D. Imhof, ChemBioChem, 2014, 15, 2754–2765.
16. P. Nockemann, K. Binnemans and K. Driesen, Chem. Phys. Lett., 2005, 415, 131–135.
17. P. Domínguez de María, J. Chem. Technol. Biotechnol., 2014, 89, 11–18.
18. Z. J. Chen, H. W. Xi, K. H. Lim and J. M. Lee, Angew. Chem., Int. Ed., 2013, 52, 13392–13396.
19. W. L. F. Armarego and D. D. Perrin, The Purification of Laboratory Chemicals, Butterworth-Heinemann, Oxford, 4th edn, 1997.
20. R. Mackenzie and M. Ritchie, Proc. R. Soc. A, 1946, 185, 207–224.
21. Y. Noritake, N. Umezawa, N. Kato and T. Higuchi, Inorg. Chem., 2013, 52, 3653–3662.
22. I. AlNashef, M. Hashim, F. Mjalli, M. Q. Al-haj Ali and M. Hayyan, Tetrahedron Lett., 2010, 51, 1976–1978.
23. C. Domínguez, M. Munoz, A. Quintanilla, Z. M. de Pedro, S. P. M. Ventura, J. A. P. Coutinho, J. A. Casas and J. Rodriguez, J. Chem. Technol. Biotechnol., 2014, 89, 1197–1202.
24. A. J. Arduengo, R. L. Harlow and M. Kline, J. Am. Chem. Soc., 1991, 113, 361–363.
25. J. Mendez-Hurtado, R. Lopez, D. Suarez and M. I. Menendez, Chem.–Eur. J., 2012, 18, 8437–8447.
26. A. Taheri, B. Lai, C. Cheng and Y. Gu, Green Chem., 2015, 17, 812–816.
27. B. Puértolas, A. K. Hill, T. Garcia, B. Solsona and L. Torrente-Murciano, Catal. Today, 2014 DOI:10.1016/j.cattod.2014.03.054.
28. V. M. Kobryanksii and S. A. Arnautov, Synth. Met., 1993, 55, 924.
29. A. P. Doherty and C. A. Brooks, Electrochim. Acta, 2004, 49, 3821.
30. M. Chi-Yung, K. Y. Wong and T. H. Chan, Chem. Commun., 2005, 1345–1347.

---

Footnote

† Electronic supplementary information (ESI) available: Materials/methods, NMR spectra, dosage of H₂O₂, UV-vis spectra. See DOI: 10.1039/c5ra01080b

---