Influence of gamma irradiation on hydrophobic room-temperature ionic liquids [BuMeIm]PF₆ and [BuMeIm](CF₃SO₂)₂N

Abstract

Stability of neat hydrophobic Room-Temperature Ionic Liquids (RTIL) [BuMeIm]X, where [BuMeIm]⁺ is 1-butyl-3-methylimidazolium and X⁻ is PF₆⁻, and (CF₃SO₂)₂N⁻, was studied under gamma radiolysis (¹³⁷Cs) in an argon atmosphere and in air. It was found that the density, surface tension, and refraction index of RTILs are unchanged even by an absorbed dose of approximately 600 kGy. Studied RTILs exhibit considerable darkening when subjected to gamma irradiation. The light absorbance of ionic liquids increases linearly with the irradiation dose. Water has no influence on radiolytic darkening. A comparative study of [BuMeIm]X and [Bu₄N][Tf₂N] leads to the conclusion that the formation of colored products is related to gamma radiolysis of the [BuMeIm]⁺ cation. The radiolytic darkening kinetics of RTILs is influenced by the anions as follows: Cl⁻ < (CF₃SO₂)₂N⁻ < PF₆⁻. Electrospray ionization mass spectrometry and NMR analysis reveal the presence of nonvolatile radiolysis products at concentrations below 1 mol% for an absorbed dose exceeding 1200 kGy. Initial step of BuMeIm⁺ cation radiolysis is the loss of the Bu˙ group, the H˙ atom from the 2 position on the imidazolium ring, and the H˙ atom from the butyl chain. Radiolysis of ionic liquid anions yields F˙ and CF₃˙ from PF₆⁻ and [Tf₂N]⁻, respectively. Recombinations of these primary products of radiolysis lead to various polymeric and acidic species.

---

Introduction

Water-stable room-temperature ionic liquids (RTIL) containing organic cations, for instance 1,3-dialkylimidazolium, or tetraalkylammonium, associated with various inorganic anions such as hexafluorophosphate (PF₆⁻), or bis(trifluoromethylsulfonyl)imide ((CF₃SO₂)₂N⁻, or [Tf₂N]⁻), are considered as promising solvents for chemical processing due to their negligible vapor pressure, good conductivity, wide electrochemical window (>4 V), and high thermal stability.¹ The unique properties of ionic liquids make them attractive for future processes in the nuclear fuel cycle including actinide electrorefining² and solvent extraction.³ Recently, we have published studies on actinide behavior in ionic liquids focused on spectroscopic and electrochemical investigations of AnCl₆²⁻ with An = U, Np and Pu.⁴ However, to be successful in such applications RTILs should be at least relatively robust to high radiation doses. A study of the effect of αβγ irradiation on [BuMeIm]NO₃, [EtMeIm]Cl, and [HexMeIm]Cl, where BuMeIm⁺, EtMeIm⁺, and HexMeIm⁺ are 1-butyl-3-methylimidazolium, 1-ethyl-3-methylimidazolium, and 1-hexyl-3-methylimidazolium respectively, showed that less than 1% of RTILs underwent radiolysis when exposed to a dose of 400 kGy.⁵ However, the irradiated samples exhibited considerable darkening following irradiation. The formation of colored products was attributed to carbene appearing as a result of alkylimidazolium radiolysis. The primary effects of radiation in RTILs were studied using pulse radiolysis techniques.⁶ It was shown that the electrons produced by ionization are very rapidly captured by alkylimidazolium cation followed by the formation of imidazoyl radicals. Secondary reactions of these radicals may lead to the formation of carbene and imidazoyl dimerisation products. In tetraalkylammonium-based RTILs solvated electrons react much more slowly with the solvent. Radiation-induced fragmentation of [Tf₂N]⁻ produces CF₃˙ radicals through dissociation of intermediate Tf₂N˙ “hole” radicals or excited [Tf₂N]⁻* anions.⁷ Moreover, it is well known that radiation products may have considerable influence on the physical properties of solvents, although no data on this topic has been published in the literature for RTILs.

The aim of the present work was to investigate the composition of [BuMeIm]⁺ gamma radiation products and the effect of gamma irradiation on the physical properties of neat hydrophobic ionic liquids [BuMeIm]PF₆ and [BuMeIm][Tf₂N], where [BuMeIm]⁺ is 1-butyl-3-methylimidazolium and [Tf₂N]⁻ is bis(trifluoromethylsulfonyl)imide. RTILs were exposed to a total radiation dose ranged from 100 to 600 kGy, and also 1200 kGy (dose rate 32–34 Gy min⁻¹) in flowing argon or simply left open to air.

Experimental

Chemical reagents

Ionic liquids with [BuMeIm]⁺ cation were prepared by a metathesis reaction from [BuMeIm]Cl and corresponding HX (X⁻ = [Tf₂N]⁻ and PF₆⁻). [BuMeIm]Cl was synthesized via quaternisation of 1-methylimidazole by 1-chlorobutane, washed with ethyl acetate and dried at reduced pressure as recently described.³ [Bu₄N][Tf₂N] was prepared from [Bu₄N]OH and Tf₂NH as in the reference.⁶ RTILs were washed with deionized water (18 MΩ cm) to neutral reaction. The absence of chloride ions was controlled by an AgNO₃ test of the aqueous phase pre-equilibrated with RTILs. Organic impurities were removed from water pre-equilibrated RTILs by mixing with activated carbon for 12 h, after which the mixture was passed through a column with small amounts of acidic alumina. The RTILs were dried in vacuo (∼5 mbar) at 80 °C for ca. 6 h. Water pre-equilibrated RTILs were prepared by mixing equal volumes of the corresponding ionic liquid and deionized water for 24 h followed by phase separation. NMR and electrospray mass spectrometry analysis revealed the absence (<0.1%) of impurities in the purified ionic liquids.

Analytical techniques

Water concentration in the RTILs was measured by coulometric Karl–Fisher titration. Spectrophotometric and ATR-IR measurements were performed using Shimadzu UV-3101 PC and Nicolet Magna-IR devices, respectively. NMR experiments were carried out with a Varian-INOVA 400. A triple resonance indirect detection probe is used for ¹⁵N and ¹³C–¹H long range correlation experiments. All other nucleus are directly observed with a broad band probe. Samples were diluted in anhydrous DMSO (Aldrich ref. 15, 187-4) and the chemical shift scale was calibrated with DMSO (2.42 ppm and 40.97 ppm) for ¹H and ¹³C, respectively, CH₃NO₂ (0 ppm) for ¹⁵N, C₆F₆ (−162.73 ppm) for ¹⁹F and H₃PO₄ (83%) (0 ppm) for ³¹P. The mass spectrometry measurements were recorded in positive and negative mode using a Bruker Esquire-LC quadrupole ion trap mass spectrometer equipped with an ESI interface. Mass spectra were acquired over a mass range of m/z from 45 to 2200. The nebulizing gas pressure (N₂) was set to 5 psi. Nitrogen drying gas was used at a flow rate of 5 L min⁻¹ at 250 °C. Ion spray voltage, cap exit offset, and trap drive were 4000, 60 and 50 V, respectively, in positive mode (ESI⁺) and −4000, −60 and −50 V, respectively, in negative mode (ESI⁻). Skimmer 1 voltage was set within the range 20–80 V, skimmer 2 voltage was kept constant at 10 V. Samples were diluted (1/10 000) in acetonitrile–water (1 : 1) mixture and were infused using a syringe infusion pump (Cole Parmer) at a flow rate of 90 µL h⁻¹. For MSⁿ experiments, helium was used as collision gas.

Before radiolysis, the analytical characteristics of the RTILs are:

[BuMeIm]Cl. ¹H NMR: 0.97 (t), 1.38 (sextet), 1.91 (qnt), 4.13 (s), 4.34 (t), 7.46 (t), 7.60 (t) and 10.66 (s). ESI⁺ (skimmer 1 +30 V): 139.1 (C₈H₁₅N₂⁺, 100%), 83.2 (C₄H₇N₂⁺, 5.5%), 313.1 ([M + C₈H₁₅N₂]⁺, <1%). ESI^– (skimmer 1 −30 V): 383.0 ([M₂Cl]⁻, 100%), 558.9 ([M₃Cl]⁻, 51%), 732.9 ([M₄Cl]⁻, 13%), 906.6 ([M₅Cl]⁻, <3%); M = [BuMeIm]Cl.

[BuMeIm][Tf₂N]. ¹H NMR: 0.83 (t), 1.21 (sextet), 1.71 (qnt), 3.77 (s), 4.08 (t), 7.56 (t), 7.63 (t) and 8.97 (s). ¹³C NMR: 13.19, 19.33, 31.92, 36.33, 49.97, 113.41, 117.67, 121.93, 122.23, 123.65, 126.19 and 136.14. ¹⁵N NMR: −1.3 and −13. ¹⁵N nucleus of the [Tf₂N]⁻ is not observed because there is no ¹H at the nitrogen vicinity. ¹⁹F NMR: −78.60. IR (cm⁻¹) (ATR): 3107, 3122, 3157 (ν_C–H str. aromatic), 2950, 2880, 2968 (ν_C–H str. aliphatic), 1573, 1466 (ring str. sym.), 1431 (MeC–H sym). ESI⁺ (skimmer 1 +30 V), 139.2 (C₈H₁₅N₂⁺, 100%), 83.2 (C₄H₇N₂⁺, 3.6%), 558.0 ([M + C₈H₁₅N₂]⁺, 0.6%). ESI^– (skimmer 1 −30 V), 279.9 (C₂O₄NS₂F₆⁻, 100%), 698.6 ([M + C₂O₄NS₂F₆]⁻, <5%), 148.0 (CO₂NSF₃⁻, <1%); M = [BuMeIm][Tf₂N].

[BuMeIm]PF₆. ¹H NMR: 0.72 (t), 1.15 (sextet), 1.68 (qnt), 3.73 (s), 4.05 (t), 7.22 (t), 7.30 (t) and 8.26 (s). ¹³C NMR: 17.72, 24.03, 36.52, 40.58, 54.41, 127.16, 128.49 and 141.27. ¹⁵N NMR: −1.3 and −13. ¹⁹F NMR: −70.03 (d) ¹J_PF = 711.2 Hz. ³¹P NMR: 42.17 (septet ¹J_PF = 711.2 Hz). IR (cm⁻¹) (ATR): 3171, 3152 (ν_C–H str. aromatic), 2966, 2936, 2878 (ν_C–H str. aliphatic), 1575, 1467 (ring str. sym.), 1432, 1386 (MeC–H sym), 1339, 1168 (ring str sym). ESI⁺ (skimmer 1 + 30 V): 139.1 (C₈H₁₅N₂⁺, 100%), 83.2 (C₄H₇N₂⁺, 5.5%), 423.4 ([M + C₈H₁₅N₂]⁺, <1%). ESI⁻ (skimmer 1 −30 V): 144.8 (PF₆⁻, 100%), 428.9 ([M + PF₆]⁻, 0.1%); M = [BuMeIm]PF₆.

Physicochemical measurements

Density, viscosity, conductivity, surface tension, and refraction index were measured at 25 °C ± 0.5 °C with a DMA 4500 density meter, Ostwald capillary viscometer, Consort K320 conductometer, KRUSS K10 tensiometer (Pt sheet), and OPL refractometer, respectively. The DSC/TGA measurements were made using an STA 409C instrument (5 K min⁻¹, aluminium oxide crucible, heating in air).

The physical properties of the RTILs before radiolysis are shown in Table 1. Density, refraction index, and the thermal decomposition temperature are in good agreement with the data published in the literature, and the presence of water has little effect on these properties.

Table 1 Physical properties of RTILs at 25 °C ± 0.5 °C

Solvent	Density/g cm^−3	Surface tension/dyn cm^−1	Refraction index	Conductivity/mS cm^−1	Viscosity/cP	Thermal decomposition^g/°C
a [H2O] = 0.027 M.b [H2O] = 0.83 M.c [H2O] = 0.035 M.d [H2O] = 0.15 M.e RTIL saturated with water.f Data at 20 °C.g For 10% lost mass.
[BuMeIm][Tf2N] ([H2O] = 0.022 M)	1.435	35.4	1.426	6.5	45.1	432
	1.43 [8]	48.8 [8]			69 [8]	439 [8]
	1.429^f [9]		1.427^f [9]	4.9^f [9]	52^f [9]
	1.39^e [8]	49.8^e [8]			62.4 [13]^a
					42.6 [13]^b
[BuMeIm]PF6 ([H2O] = 0.038 M)	1.366	46.8	1.408	2.45	235	391
	1.36 [8]	37.5 [8]	1.409 [8]	1.46 [11]	375 [13]^c	394 [8]
		49.8^f [10]		1.8 [12]
[BuMeIm]PF6·H2O ([H2O] = 0.840 M)	1.362	46.3	1.408	3.34	137	360 [8]
	1.35 [8]	36.8 [8]			397 [8]
					328 [13]^d

---

Concerning viscosity, conductivity and surface tension, our measurements are somewhat different from those reported in the literature. However, it was noted that the presence of water decreases RTIL viscosities^8,13 leading to an increase in their conductivities. Viscosities of RTILs measured in this work and reported in Table 1, are also correlated with this general trend. As water concentrations are often unknown, it is difficult to compare our conductivity and the surface tension values with those published in the literature.

Irradiation device

Irradiation was carried out at room temperature using, a MARCEL Facility equipped with an IBL 137 Cis-Bio Irradiator (¹³⁷Cs as the gamma source dose rate 32–34 Gy min⁻¹). Dosimetry was performed using a conventional ferrioxalate dosimeter. Radiolysis in air was performed in open flasks. Argon was sparged through the flasks during the radiation tests in an inert atmosphere.

Results

Radiolysis under argon atmosphere

Argon flow prevents absorption by RTILs of water and oxygen from the atmosphere during radiolysis. Analysis by the Karl–Fisher method revealed that water concentration in [BuMeIm][Tf₂N] increased slowly from 0.02 M to 0.08 M during 300 h of radiolysis in the presence of argon. Gamma radiolysis of 30% TBP–dodecane, [BuMeIm]Cl (mp 54 °C) and [Bu₄N][Tf₂N] (mp 85–86 °C) was also tested under the same conditions for comparison with [BuMeIm][Tf₂N] and [BuMeIm]PF₆. A comparative study with 30% TBP–dodecane system is of interest since today this is a major solvent for spent nuclear fuel reprocessing.¹⁴

The evolution of the physical properties of RTILs during radiolysis in argon is shown in Table 2. Radiolysis has no significant impact on the density, surface tension, or refraction index of RTILs. The density variation of irradiated 30% TBP–dodecane solvent is higher than for RTILs. However, Table 2 shows that radiolysis in argon atmosphere increases the viscosity of RTILs twice more than for 30% TBP–dodecane at a comparable absorbed dose. It should be noticed that even relatively small amounts of impurities can have dramatic impact on ionic liquid viscosity.¹⁵ The higher viscosity is probably related to hydrogen bonding between the radiolysis products and the protons of the imidazolium ring. Increased [BuMeIm]BF₄ and [BuMeIm]PF₆ viscosities were also observed in the presence of chloride ions which are known to form hydrogen bonds with BuMeIm⁺ cation.¹⁵ It can thus be assumed that radiolysis of studied RTILs leads to an accumulation of polar products able to create strong interactions with alkylimidazolium cations. Table 2 also shows that conductivity of RTILs decreases during radiolysis. This effect can be attributed to the increase in RTIL viscosity under gamma radiation. Washing irradiated RTILs with deionized water yields an aqueous phase at pH ≈ 2–3 indicating that acids are formed during irradiation. These acidic radiolysis products may increase the RTIL conductivity, although their concentration is probably too low to have a considerable impact on solvent conductivity. For [BuMeIm]PF₆ the acidic products, most probably, are related to HF formed as a result of PF₆⁻ degradation by similarity with well known PF₆⁻ hydrolysis.¹⁶

Table 2 Evolution of the physical properties of RTILs (at 25 °C ± 0.5 °C) during γ-radiation under argon atmosphere

Solvent	Absorbed dose/kGy	Water/M	Density/g cm^−3	Surface tension/dyn cm^−1	Refraction index	Conductivity/mS cm^−1	Viscosity/cP
[BuMeIm][Tf2N]	0	0.022	1.44	35.4	1.426	6.59	45.1
	185	—	1.44	35.0	1.427	6.60	47.5
	525	0.080	1.44	34.6	1.429	5.30	63.4
[BuMeIm]PF6	0	0.038	1.36	46.8	1.408	2.45	235
	579	—	1.37	45.0	1.410	1.96	354
30%TBP–dodecane	0	—	0.82	25.7	1.420	—	1.79
	201	—	0.83	26.4	1.420	—	1.87
	572	—	0.86	26.5	1.421	—	2.14

---

Electrospray ionization mass spectrometry and NMR analysis of 580 kGy irradiated [BuMeIm]PF₆ and [BuMeIm][Tf₂N] samples did not show the presence of degradation products in significant quantities. From this observation, it can be concluded that a very small amount of degradation products (<0.5 %mol) is formed.

[BuMeIm][Tf₂N], [BuMeIm]PF₆, and [BuMeIm]Cl show darkening during γ-radiolysis similar to that observed for alkylmethylimidazolium ionic liquids with nitrate or chloride anion.⁵ Typical UV/VIS spectra of irradiated RTILs are shown in Fig. 1. Gamma irradiation leads to appearance of broad light absorption bands to appear with a maximum of less than 300 nm. By contrast, radiolytic darkening of [Bu₄N][Tf₂N] and 30% TBP–dodecane is significantly less intensive than that of BuMeIm⁺ based ionic liquids. Fig. 2 demonstrates that the light absorbance of RTILs at λ = 340 nm increases linearly with the irradiation dose, suggesting that colored products come from radiolysis of the organic cation. The rate of radiolytic darkening varies in the following order: [BuMeIm]Cl < [BuMeIm][Tf₂N] < [BuMeIm]PF₆. The thermogravimetric measurements reported in Table 1 are consistent with those published by Huddleston et al.⁸ and confirm the stability order for RTILs constituted by the [BuMeIm]⁺ cation with different anions. These results clearly indicate that the kinetics of radiolytic darkening is influenced by the ionic liquid anions.

Fig. 1 UV/VIS spectra of [BuMeIm][Tf₂N] before (1) and after γ-irradiation at 270 kGy (2) and 450 kGy (3) under argon atmosphere.

Fig. 2 Dependence of light absorbance on γ-dose on UV/VIS spectra of irradiated RTILs at 340 nm under argon atmosphere, [BuMeIm]PF₆ (1), [BuMeIm][Tf₂N] (2), [BuMeIm]Cl (3), [Bu₄N][Tf₂N] (4) and 30%TBP–dodecane (5), l = 2 mm. Samples were diluted 100 times with acetone before measurements.

Radiolysis in air

Analysis by the Karl–Fisher method indicate that dried samples of irradiated RTILs exposed to air during γ-radiolysis (300 h) absorb water as shown in Table 3 due to the hygroscopic properties of RTILs. ATR-IR spectra of irradiated RTILs reveal the bands at ∼3600 cm⁻¹ and ∼1660 cm⁻¹ related to O–H stretch and H–O–H deformation vibrations, respectively, of absorbed H₂O molecules.

Table 3 Concentration of water in RTILs before and after 228 h (464 kGy) of radiolysis under air atmosphere

Solvents	[H2O]/M
	Before radiolysis	After radiolysis
a [BuMeIm]PF6 pre-equilibrated with water.
[BuMeIm][Tf2N]	0.022	0.470
[BuMeIm]PF6	0.038	0.460
[BuMeIm]PF6·H2O^a	0.840	0.740

---

All samples show darkening following radiolysis similar to that observed in an argon atmosphere. The kinetics of radiolytic darkening in air exhibit the same behavior as in argon. Fig. 3 demonstrates the linear increase of RTIL light absorbance with the irradiation dose. The color change is more rapid for [BuMeIm]PF₆ than that for [BuMeIm][Tf₂N]. The presence of water has no significant influence on the darkening of [BuMeIm]PF₆ during radiolysis.

Fig. 3 Dependence of light absorbance on gamma-dose on UV/VIS spectra of irradiated RTILs at a wavelength of 464 nm under air atmosphere, [BuMeIm][Tf₂N] (1), [BuMeIm]PF₆ (2) and [BuMeIm]PF₆·H₂O (3). l = 2 mm.

Data on the physical properties of irradiated RTILs presented in Table 4 indicate that, as in argon, gamma radiolysis in air has no significant effect on the densities, surface tensions, and refraction index of RTILs. The observed changes in viscosity during radiolysis are probably related to two opposite factors: (i) decrease of RTIL viscosity due to water absorption, (ii) increase of RTIL viscosity as a result of radiolysis. The effect of water on RTIL viscosity is opposite to that of the radiolytic products, probably, due to the difference in their bonding with ionic liquids. Water is known to bind via H-bonding with the anions of ionic liquids.¹⁷ By contrast, the radiolytic products probably interact with the cation. The changes in conductivity of irradiated RTILs follow the variation of their viscosities: conductivities decrease when the viscosities rise.

Table 4 Evolution of the physical properties of RTILs (at 25 °C ± 0.5 °C) during γ-radiation under air atmosphere

RTIL	Dose/kGy	Density/g cm^−3	Surface tension/dyn cm^−1	Refraction index	Conductivity/mS cm^−1	Viscosity/cP
[BuMeIm][Tf2N]	0	1.44	35.4	1.426	6.59	45.1
	284	1.43	35.3	1.426	5.61	52.5
	464	—	35.5	1.426	4.57	—
[BuMeIm]PF6	0	1.36	46.8	1.408	2.45	235
	272	1.36	46.9	1.408	2.70	189
	444	—	47.1	1.409	2.56	—
[BuMeIm]PF6·H2O	0	1.36	46.3	1.408	3.34	137
	278	1.36	46.6	1.409	2.83	186
	453	—	46.9	1.409	2.77	—

---

No changes were observed in the NMR and electrospray spectra of RTILs recorded before and after irradiation at 450 kGy, suggesting that less than 0.5% of ionic liquids underwent radiolysis.

Identification of radiolysis products

In order to identify the more stable degradation products, RTILs were exposed to a total radiation dose of about 1.2 MGy (dose rate 32–34 Gy min⁻¹) in an argon flow. These solutions are characterized by NMR and by electrospray ionization (ESI⁺ and ESI⁻) mass spectrometry. Electrospray ionization mass spectrometry has been shown to be an appropriate technique to characterize qualitatively preformed ions in solution. The MSⁿ spectra of degradation products give information about the structure of these compounds. It should also be noticed that the specific odor of irradiated solvents clearly indicates the presence of volatile sulfur-containing radiolytic products. However, the gaseous products cannot be identified by the analytical methods used in this work.

Mass spectrometry and NMR studies were undertaken to verify the absence of impurities in the RTILs (see analytical section) and to identify degradation products after radiolysis: (i) cations with higher m/z ratios due to the potential formation of carbene and (ii) new compounds with lower m/z ratios due to the degradation of the initial molecule.

Before the qualitative study of degraded solution, the effect of the skimmer 1 voltage was studied in particular on positive (ESI⁺) and negative (ESI⁻) ionization mass spectra of [BuMeIm][Tf₂N] non irradiated. The MS² spectra of the main ion were recorded to yield information about the fragmentation pathway in the gas phase. The positive ionization mass spectra (ESI,† Fig. S1a) show a peak at m/z = 139.1 corresponding to [BuMeIm]⁺ regardless of the skimmer 1 voltage. Additional peaks were observed depending on the skimmer voltage. For skimmer 1 voltages below 30 V, dimeric ions M[BuMeIm]⁺ (with M = [BuMeIm][Tf₂N]) (m/z = 558.0) were observed showing the molecular association between molecules in the gas phase. Increasing skimmer 1 voltage led to the disappearance of dimeric ions and fragmentation of the [BuMeIm]⁺ cation. A fragment at m/z = 83.2 appeared to lose a butene group, its intensity increased with the skimmer 1 voltage. Moreover, the MS² spectra of [BuMeIm]⁺ show the same fragmentation pathway: the loss of a butene molecule (ESI,† Fig. S2). The negative ionization mass spectra (ESI,† Fig. S1b) show a peak at m/z = 279.9 corresponding to [Tf₂N]⁻ regardless of the skimmer voltage. Dimeric ions M[Tf₂N]⁻ (with M = [BuMeIm][Tf₂N]) (m/z = 698.6) were observed at skimmer 1 voltages below 30 V showing, as in positive ionization mode, the molecular association between molecules in the gas phase. Increasing skimmer 1 voltage (above 30 V) led to the disappearance of dimeric ions and fragmentation of [Tf₂N]⁻ anion. A fragment at m/z = 148.0 appeared due to loss of a SO₂CF₃ group. Its intensity increases with skimmer voltage. Moreover, the MS² spectra of the [Tf₂N]⁻ show the loss of a CF₃ group and a SO₂CF₃ group (ESI,† Fig. S3). These results obtained by electrospray ionization (ESI⁺ and ESI⁻) are consistent with those published by Alfassi et al.^18a and Jackson et al.^18b

For comparison with [BuMeIm][Tf₂N], electrospray ionization mass spectra of [BuMeIm]PF₆ and [BuMeIm]Cl before radiolysis were recorded in positive (ESI⁺) and negative ionization (ESI⁻) mode (results not shown). For example, ESI,† Fig. S4 shows in negative mode the formation of M_nCl⁻ supramolecular structures (where M = [BuMeIm]Cl) with n = 2–6 (m/z = 383.0, 558.9, 732.9, 906.6 and 1080.7 respectively) in the gas phase. In the case of [BuMeIm]BF₄, Dupont¹⁹ has observed this supramolecular structure in the gas phase. These preliminary studies show the importance of the experimental conditions when analyzing RTILs with optimum sensitivity without molecular association and fragmentations.

Fig. 4 (a) ESI(+) mass spectra of the [BuMeIm]PF₆ after radiolysis at 1.2 MGy in acetonitrile–water solution. Skimmer 1 voltage: 30V. (b) ESI(+) mass spectra of the [BuMeIm][Tf₂N] after radiolysis at 1.2 MGy in acetonitrile–water solution. Skimmer 1 voltage: 30 V.

As an example, Fig. 4 shows the electrospray ionization mass spectrum in positive ionization mode of [BuMeIm]X (X⁻ = [Tf₂N]⁻ and PF₆⁻) after radiolysis (1200 kGy). For both ionic liquids, this spectrum shows:

(i) a peak at m/z = 83.2, with 40% intensity for X⁻ = [Tf₂N]⁻ and 36% for X⁻ = PF₆⁻, compared with 4% and 6%, respectively, before radiolysis;

(ii) two very low-intensity peaks: at m/z = 207.1 for X⁻ = [Tf₂N]⁻ and m/z = 157.1 for X⁻ = PF₆⁻;

(iii) two pairs of very low intensity peaks: one pair of which corresponds to the dimeric ions already observed before radiolysis with a low skimmer voltage (m/z = 558.0 and m/z = 423.4), and another pair of peaks at m/z = 556.2 for X⁻ = [Tf₂N]⁻ and m/z = 421.3 for X⁻ = PF₆⁻.

The intensity ratio of the peak at m/z = 83.2 versus the peak of BuMeIm⁺ (at m/z = 139.1) is 10 to 6 times higher after radiolysis for X⁻ = [Tf₂N]⁻ and PF₆⁻ respectively, compared with the value of the initial solution. This result shows that rupture of the butyl chain forms one of the radiolysis products of the imidazolium cation.

The peaks (Fig. 4) at m/z = 207.1 and m/z = 157.1 were assigned to a [BuMeIm–Y]⁺ cation where Y = CF₃ in the case of the [Tf₂N]⁻ anion and Y = F in the case of the PF₆⁻ anion on the basis of the MS² spectra. The MS² spectra of these two cations show the elimination of a butyl chain resulting in two peaks at m/z = 151.0 for X⁻ = [Tf₂N]⁻ and m/z = 101.0 for X⁻ = PF₆⁻. Moreover, in the case of X⁻ = PF₆⁻ it was observed that the loss of HF resulted in an ion at m/z = 137.1.

Also, a new low intensity peak is observed corresponding to a [(BuMeIm)₂X]⁺ structure at m/z = 556.2 (for X⁻ = [Tf₂N]⁻) and at m/z = 421.3 (for X⁻ = PF₆⁻). Fragmentations of these two cations (MS²) (Fig. 5) lead to the main ions at m/z = 275.2 (loss of HX with X⁻ = PF₆⁻ or X⁻ = [Tf₂N]⁻), and m/z = 193.2 (loss of HX followed by the loss of C₄H₆N₂). In both cases, the ions at m/z = 275.2 have been isolated and fragmented. The MS³ spectra (Fig. 6) show that the structures of these 2 ions are different and depend on their origin. In the case of the [Tf₂N]⁻ anion, the MS³ spectra of the ions at m/z = 275.3, arising from the ion at m/z = 556.3, show an ion at m/z = 219.2 due to the loss of the butene group and an ion at m/z = 193.2 due to the loss of C₄H₆N₂. In the case of PF₆⁻ anion, the MS³ spectra of the ions at m/z = 275.2, arising from the ion at m/z = 421.3, show only an ion at m/z = 193.2 due to a loss of C₄H₆N₂.

Fig. 5 (a) MS² spectra of ion at m/z = 421.3 after radiolysis of [BuMeIm]PF₆ at 1.2 MGy. (b) MS² spectra of ion at m/z = 556.3 after radiolysis of [BuMeIm][Tf₂N] at 1.2 MGy.

Fig. 6 (a) MS³ spectra of ion at m/z = 275.2 after radiolysis of [BuMeIm]PF₆ at 1.2 MGy, issuing of MS² spectra of the ion at m/z = 421.3. (b) MS³ spectra of ion at m/z = 275.2 after radiolysis of [BuMeIm][Tf₂N] at 1.2 MGy, issuing of MS² spectra of the ion at m/z = 556.3.

The ESI mass spectra show that no further peaks are observed in the negative ionization mode, regardless of the anion (PF₆⁻ or [Tf₂N]⁻).

The ¹H NMR study of these two RTILs after irradiation (1200 kGy) shows very low intensity signals (<0.5%) which are observed between 0 and 5 ppm after eliminating the ¹³C coupling signals. Because of very low intensity and poor resolution of these signals, it was impossible to identify them. Moreover, these products could not be identified by selective irradiation experiments.

The ¹⁹F NMR spectrum of the ionic liquid comprising the [Tf₂N]⁻ anion revealed a singlet at −79.28 ppm (¹J_CF = 321 Hz) with an intensity of about 1.5% of the ¹⁹F signal, and several peak clusters with intensities below 0.05% of the signal between −76 ppm and −81 ppm. Although it is difficult to analyze very low-intensity peaks, the singlet at −79.28 ppm probably corresponds to fluorine in a CF₃SO₂N group (CF₃SO₂N–Gr). NMR signatures of the reference compounds (chemical shift and coupling constants) measured under the same conditions (e.g. (CF₃SO₂)₂NH: −77.60 ppm (¹J_CF = 322 Hz), CF₃SO₃H: −77.84 ppm (¹J_CF = 321.4 Hz) and CF₃SO₂(N₂C₃H₃): −76.19 (¹J_CF = 323.0 Hz), unlike CF₃CH₂OH: −75.40 ppm (¹J_CF = 279.2 Hz and ²J_HF = 9.5 Hz)) are very close to those of the singlet at −79.28 ppm. Moreover, HCF₃, which is liable to be formed by radiolysis of [Tf₂N]⁻ anion, was not observed by NMR analysis of ¹H nuclei (quartet at 6.25 ppm; ²J_HF = 79 Hz from ¹H spectrum calculation) or ¹⁹F nuclei (doublet near 0 ppm) probably due to the high volatility of this compound.

The ¹⁹F and ³¹P NMR spectra of the ionic liquid consisting of PF₆⁻ anion revealed a PF₂Z chemical function, with the signature: ¹⁹F, a doublet at −77.22 ppm and −79.75 ppm (¹J_PF = 950 Hz) and ³¹P, a triplet at 171.4 ppm (¹J_PF = 950 Hz) with an intensity of about 0.65% of the phosphorus signal. The absence of coupling with ¹H and ¹³C suggests that Z contains neither carbon nor hydrogen. It is possible that this compound is formed following the rupture of the P–F bonds, possibly leading to compounds such as fluorophosphate (PO_xF₂) in contact with traces of water. It should also be noticed that the characteristic signal of HF (at −167.32 ppm; ¹J_HF = 392.5 Hz) was not observed after irradiation at 1200 kGy. This can be related to high volatility of this acid and relatively low sensitivity of NMR analysis.

Discussion

The results above show the small amount of products formed following radiolysis of the two ionic liquids. Clearly the physical and chemical properties of the two liquids are only slightly affected. However, the modifications of the conductivity—a property particularly sensitive to the presence of solute in solution—show that radiolysis products accumulate in solution. The evolution of the solution coloration with the absorbed dose also confirms the solute accumulation during radiolysis.

Concerning the reaction mechanism, results are difficult to interpret because of the small amount of radiolysis products and the large number of compounds formed. Nevertheless, the initial results suggest that the primary products of RTILs radiolysis are the following:

(i) the methylimidazolium cation radical [MeIm˙]⁺ arising from rupture of the butyl chain on the nitrogen atom of the imidazolium ring),

(ii) [BuMeIm˙]⁺ cation radical formed after the homolytical dissociation of the H–C2 bond of the imidazolium ring,

(iii) [Bu˙MeIm]⁺ cation radical formed following the loss of a hydrogen atom from the butyl chain,

(iv) CF₃˙ and F˙ radicals from [Tf₂N]⁻ and PF₆⁻ anions, respectively.

The secondary reactions of these radicals yield numerous stable products of radiolysis at very low concentrations. Formation of the following major species can be assumed (Schemes 1 and 2):

Scheme 1 Proposed simplified degradation scheme of [BuMeIm][Tf₂N].

Scheme 2 Proposed simplified degradation scheme of [BuMeIm]PF₆.

(i) [BuMeIm–Y]⁺ cation, where Y = CF₃ and F for [BuMeIm][Tf₂N] and [BuMeIm][PF₆], respectively, formed by recombination of the primary radicals produced by radiolysis of the anion (CF₃˙ and F˙) and [BuMeIm˙]⁺ cation radical,

(ii) [(BuMeIm)₂X]⁺ ionic pair, where [BuMeIm]₂²⁺ is a dimer of two primary cation radicals and X⁻ is [Tf₂N]⁻ or PF₆⁻. Depending on X⁻, the structure of the dimer appears to differ. In the case of [Tf₂N]⁻ the cation appears to form after recombination of a [BuMeIm˙]⁺ and a [Bu˙MeIm]⁺ cation radical (Scheme 1). In the case of PF₆⁻ the cation appears to form after recombination of two [Bu˙MeIm]⁺ cation radicals (Scheme 2).

It also appears that radiolysis of the [Tf₂N]⁻ anion forms CF₃SO₂N˙ species (diagram 1). In the case of the PF₆⁻ anion, radiolysis results in a PF₂Z species, where Z contains neither carbon nor hydrogen (diagram 2). As it was mentioned above, these species could be assigned to fluorophosphates (PO_xF₂) formed by hydrolysis of PF₅˙ primary radicals in the presence of water traces of irradiated ionic liquid.

Recombination of [MeIm˙]⁺ radicals yields the aromatic dimers with nitrogen heteroatom which are usually known to have high molar extinction coefficients in UV/VIS optical spectra. This accounts for the significant radiolytic darkening of the studied RTILs even at relatively low concentrations of the products. Obviously, the Bu₄N⁺ cation cannot form carbene or aromatic dimers and hence is unable to form colored radiolysis products.

In general, the relative radiolytical stability of studied RTILs can be interpreted in terms of the interaction of the anion ([Tf₂N]⁻ and PF₆⁻) with the cation ([BuMeIm⁺]) and anion stability. In fact, imidazolium salts are better described as polymeric supramolecules formed through hydrogen bonds of the imidazolium cation (especially on carbon C2) with the anion.²⁰ Moreover, as observed for thermal stability, the radiolytic stability of dialkylimidazolium is connected with the stability of the anion.²¹ Considering the strength of the hydrogen bonds and the stability of the anion, the relative kinetics of RTILs darkening under gamma radiolysis are PF₆⁻ > [Tf₂N]⁻ > Cl⁻.

Conclusions

An investigation of gamma radiolysis of hydrophobic ionic liquids [BuMeIm]PF₆ and [BuMeIm][Tf₂N] has shown that their densities, surface tensions and refraction indices are unchanged even after a strong absorbed radiation dose. Gamma radiolysis causes a significant increase in the viscosities of RTILs. The conductivities of the irradiated RTILs decrease as the viscosity increases.

Ionic liquids based on the butylmethylimidazolium cation exhibit considerable darkening when subjected to gamma irradiation. The light absorbance of RTILs increases linearly with the irradiation dose. Data obtained show that the formation of colored products is related to the radiolysis of the butylmethylimidazolium cation. Water has no effect on the radiolytic darkening kinetics which is influenced by the ionic liquid anion as follows: Cl⁻ < (CF₃SO₂)₂N⁻ < PF₆⁻.

Electrospray and NMR measurements indicate that the overall concentration of nonvolatile radiolysis products does not exceed 1% (mol) for RTILs irradiated up to 1200 kGy. Abstraction of the H˙ atom from the 2-carbon position of the imidazolium ring, the H˙ atom from the carbon position of the butyl chain and the Bu˙ group were found to be the primary process of the BuMeIm⁺ cation gamma radiolysis.

Acknowledgements

The authors are grateful to M. Gastaldi for assistance with radiolysis devices and CEA/DEN/DSOE and ACTINET for financial support.

References

1. Ionic Liquids in Synthesis, ed. P. Wasserscheid and T. Welton, Wiley, Weinheim, Germany, 2003.
2. (a) R. C. Thied, J. E. Hatter, K. R. Seddon, W. R. Pitner, D. W. Rooney and D. Hebditch World Pat., WO01/13379A1, 18 August 1999; (b) W. J. Oldham, D. A. Costa, Jr. and W. H. Smith, Development of Room-Temperature Ionic Liquids for Applications in Actinide Chemistry, in: Ionic Liquids. Industrial Applications for Green Chemistry, ACS Symp. Ser. 818, ed. R. D. Rogers and K. R. Seddon, Washington, DC, 2002, ch. 15, p. 188.
3. (a) A. E. Visser and R. D. Rogers, J. Solid State Chem., 2003, 109, 113 (and cited ref.); (b) S. Dai, Y. H. Ju and C. E. Barnes, J. Chem. Soc., Dalton Trans., 1999, 1201; (c) M. L. Dietz and J. A. Dzielawa, Chem. Commun., 2001, 2124.
4. (a) S. I. Nikitenko, C. Cannes, C. Le Naour, P. Moisy and D. Trubert, Inorg. Chem., 2005, 44(25), 9497; (b) S. I. Nikitenko and P. Moisy, Inorg. Chem., 2006, 45(3), 1235.
5. D. Allen, G. Baston, A. E. Bradley, T. Gorman, A. Haile, I. Hambelett, J. E. Hatter, M. J. F. Healey, B. Hodgson, R. Lewin, K. V. Lovell, B. Newton, W. R. Pitner, D. W. Rooney, D. Sanders, K. R. Seddon, H. E. Sims and R. C. Thied, Green Chem., 2002, 4, 152.
6. (a) J. F. Wishart and P. Neta, J. Phys. Chem. B, 2003, 107, 7261; (b) D. Behar, P. Neta and C. Schultheisz, J. Phys. Chem. A, 2002, 106, 3139; (c) J. Grodkowski and P. Neta, J. Phys. Chem. A, 2002, 106, 5468.
7. (a) J. F. Wishart, A. Funston and T. Szreder, XIVth Symp. On Molten Salts, The Electrochemical Society, Honolulu, Hawaii, 2004; (b) A. Funston and J. F. Wishart, Ionic Liquids IIIA: Fundamentals, Progress, Challenges and Opportunities, Properties and Structure, ACS Symp. Ser. 901, 2005, 102.7.
8. J. G. Huddleston, A. E. Visser, W. M. Reichert, H. D. Willauer, G. A. Broker and R. D. Rogers, Green Chem., 2001, 3, 156.
9. P. Bonhôte, A. P. Dias, N. Papageorgiou, K. Kalyanasundaram and M. Grätzel, Inorg. Chem., 1996, 35, 1168.
10. K. N. Marsh, J. A. Boxall and R. Lichtenthaler, Fluid Phase Equilib., 2004, 219, 93.
11. P. A. Suarez, J. E. L. Dullius, S. Einloft, R. F. de Souza and J. Dupont, Polyhedron, 1996, 15, 1217.
12. J. Fuller, A. C. Breda and R. T. Carlin, J. Electroanal. Chem., 1998, 459, 29.
13. J. A. Widegren, A. Laesecke and J. W. Magee, Chem. Commun., 2005, 1610.
14. J. J. Katz, G. T. Seaborg and L. R. Morss, The Chemistry of the Actinides, 2nd edn, Chapman & Hall, London, 1986.
15. K. R. Seddon, A. Stark and M. J. Torres, Pure Appl. Chem., 2000, 72, 2275.
16. R. P. Swatloski, J. D. Holbrey and R. D. Rogers, Green Chem., 2003, 5, 361.
17. L. Cammarata, S. G. Kazarian, P. A. Salter and T. Welton, Phys. Chem. Chem. Phys., 2001, 3, 5192.
18. (a) Z. R. Alfassi, R. E. Huie, B. L. Milman and P. Neta, Anal. Bioanal. Chem., 2003, 377, 159; (b) G. P. Jackson and D. C. Duckworth, Chem. Commun., 2004, 522.
19. J. Dupont, J. Braz. Chem. Soc., 2004, 15, 341.
20. C. S. Consorti, P. A. Z. Suarez, R. F. de Souza, R. A. Burrow, D. H. Farrar, A. J. Lough, W. Loh, L. H. M. da Silva and J. Dupont, J. Phys. Chem. B, 2005, 109, 4341.
21. J. van den Broeke, M. Stam, M. Lutz, H. Kooijman, A. L. Spek, B. J. Deelman and G. van Koren, Eur. J. Inorg. Chem., 2003, 2798.

---

Footnote

† Electronic supplementary information (ESI) available: Fig. S1a: Effect of the skimmer 1 voltage (sk1) on the ESI(+) mass spectra of [BuMeIm][Tf₂N] in acetonitrile–water solution. Fig. S1b: Effect of the skimmer 1 voltage (sk1) on the ESI(−) mass spectra of [BuMeIm][Tf₂N] in acetonitrile–water solution. Fig. S2: MS² spectra of BuMeIm⁺ (m/z = 139.1) of [BuMeIm][Tf₂N] solution in acetonitrile–water. Fig. S3: MS² spectra of Tf₂N⁻ (m/z = 279.7) of [BuMeIm][Tf₂N] solution in acetonitrile–water. Fig. S4: ESI (−) mass spectra of [BuMeIm]Cl in acetonitrile–water solution. Skimmer 1 voltage = −45 V. M = [BuMeIm]Cl. See DOI: 10.1039/b601111j

---