Reversed-phase liquid chromatography analysis of alkyl-imidazolium ionic liquids

Available online at www.sciencedirect.com

Journal of Chromatography A, 1189 (2008) 476–482

Reversed-phase liquid chromatography analysis of alkyl-imidazolium ionic liquids II. Effects of different added salts and stationary phase influence Mar´ıa Jos´e Ruiz-Angel, Alain Berthod ∗ Laboratoire des Sciences Analytiques, Universit´e de Lyon, CNRS, Bat. CPE, 69622 Villeurbanne, France Available online 22 October 2007

Abstract Two mixtures of four 1-alkyl-3-methylimidazolium ionic liquids (ILs) salts associated to the anions tetrafluoroborate or hexafluorophosphate were analyzed by reversed-phase liquid chromatography with three different stationary phases: Kromasil C8 , Zorbax Extend C18 and Zorbax Sb-Aq. The effect on retention of various inorganic salts (NaCl, NaH2 PO4, NaBF4 , NaClO4 and NaPF6 ) added to acetonitrile/water mobile phases was studied. The three columns gave similar separation profiles. In all cases, the retention of ILs increased with the increasing affinity of the inorganic anions for the apolar stationary phases; a phenomenon called chaotropicity. The chaotropic anion order is Cl− ∼ H2 PO4 − < BF4 − ∼ ClO4 − < PF6 − . It is established that the presence of chaotropic anions in the mobile phase do not permit to differentiate between ILs associated to different anions. However, chloride or dihydrogenphosphate added salts do not fully screen the retention differences between ILs associated with different anions. Distorted and even split peaks may appear in the chromatogram depending on the nature and concentration of the injected ILs. In the RPLC analysis of imidazolium-based IL, it is recommended to add to the mobile phase significant amounts of a salt containing a chaotropic anion. This salt addition will improve the IL peak shapes and give reproducible retention factors. LODs in the low microgram range (∼5 nmol) were obtained with the Kromasil C8 column with a 50/50 acetonitrile–water mobile phase containing 0.01 M NaPF6 added salt and 230 nm UV detection. © 2007 Elsevier B.V. All rights reserved. Keywords: Ionic liquids; Alkylimidazolium salts; Charge–charge interaction; Isotherm adsorption; Retention mechanisms

1. Introduction Ionic liquids (ILs) are a new class of non-molecular solvents with unique properties. They possess wide range of solubilizing capability for polar and apolar molecules as well as for ions. They have interesting electrochemical properties such as a significant dielectric constant, a high electrical conductivity and a wide electroactivity window. They also have good thermal and chemical stabilities, associated to an extremely low volatility and flammability [1,2]. The absence of vapor pressure has gained them the label of environment-friendly solvents [3]. Chemically, ILs are just room or low temperature molten salts composed by an often small inorganic anion (bromide, chloride, tetrafluoroborate, hexafluorophosphate or

∗

Corresponding author. Tel.: +33 472 431 434; fax: +33 472 431 078. E-mail address: [email protected] (A. Berthod).

0021-9673/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2007.10.046

triflate anion) and a bulky dissymmetrical organic cation, such as the dialkyl-imidazolium, alkyl-pyridinium, alkyl-ammonium or tetralkylphosphonium cation [4]. The combination of these different ions has expanded considerably the number of possible salts with low melting points that differ in properties such as viscosity, melting point or density. This variety has enormously increased the IL applications and potential in a variety of chemical processes [1–4]. In the analytical field, ILs are applied in liquid–liquid extraction and partitioning [5,6]. They are used as stationary phases in gas chromatography [7,8] and matrices in laser desorption ionization– time-of-flight-mass spectrometry (MALDI–TOFMS) [9,10]. In capillary electrophoresis, ILs are used as running electrolytes [11,12] and in reversed-phase liquid chromatography, they reduce the silanol effects when employed as mobile phase additives [13–15]. Recent synthesis of chiral ILs has opened the way of the evaluation of new potential selectors for enantiomeric separations [16,17]. In other chro-

M.J. Ruiz-Angel, A. Berthod / J. Chromatogr. A 1189 (2008) 476–482

matographic modes, such as countercurrent chromatography, ILs have been employed as stationary phases [18,19]. Several reports have reviewed their application in analytical chemistry [20–22]. The unique properties of ILs combined with their potential uses have increased the interest of researchers and industries alike. The need to develop analytical procedures for the analysis and control of ILs is not far and must be possible. Some recent reports have focused on the development of simple analytical methods that make use of separation techniques with success [23,24]. Reversed-phase liquid chromatography (RPLC) seems to be the most common technique. Nevertheless, due to the nonmolecular nature of ILs, the RPLC retention mechanism of ILs is not as simple as the one for molecular species, and specific investigations are needed. In a recent report, we clearly reminded that, in mobile phase solutions, ILs are just dissociated salts with cations and anions [25]. In the first article of this series [26], the chromatographic behavior of eleven alkyl-imidazolium ILs was investigated with different acetonitrile/water mobile phases using a Kromasil C18 column. We identified a complex RPLC retention mechanism involving ionic and hydrophobic interactions [26]. Under reversed-phase conditions with salt-free mobile phases, imidazolium-based ILs could be differentiated by the anion and/or by the cation. Mobile phases containing added salts could only differentiate ILs with different alkyl chains on the imidazolium ring. These salt containing mobile phases could no longer separate ILs differing only by their anions [25]. However, from a quantitative analysis point of view, salt containing mobile phases gave much better analytical figures of merit than the similar mobile phases without added salt [26]. The present work proposes to expand this study working with different stationary phases and salt additives in order to confirm and/or to improve the knowledge on the separation mechanism and the IL quantitative analysis. 2. Experimental

477

methyl imidazolium tetrafluoroborate or hexafluorophosphate. They were purchased from Solvent Innovation (K¨oln, Germany) except EMIM and HMIM BF4 which were from Fluka (Buchs, Switzerland). They were used as received without further purification (98% purity or more). Mobile phases were prepared with acetonitrile (SDS, Carlo Erba Reagents, Peypin, France) and distilled water. The following inorganic salts were added: sodium chloride, sodium tetrafluoroborate, (both from Fluka), sodium dihydrogenphosphate, sodium perchlorate and sodium hexafluorophosphate (all from Sigma–Aldrich, L’Isles d’Abeau Chesnes, France). Stock standard solutions (∼0.05 M) of all the salts were prepared in acetonitrile/water mixtures, and work solutions were conveniently diluted with the same aqueous–organic solution. 2.2. Stationary phases Three different columns were used. Their physico chemical and chromatographic characteristics are as follows: a 150 mm × 4 mm I.D. Kromasil C8 column was obtained from Cluzeau Infolabo, Bordeaux, France. The Kromasil C8 packing has a 5 ␮m silica particle diameter with 330 m2 /g surface area and 11 nm pore diameter. The 9% carbon content allows calculating a 2.5 ␮mol/m2 of C8 monomeric bonding density. The two other columns were from Agilent, Palo Alto, USA. A 150 mm × 3 mm I.D. Zorbax Extend C18 contained a packing made with 5 ␮m ultra pure, metal free silica particle with 180 m2 /g surface area and 8 nm pore diameter. A special proprietary bidentate C18 moiety was attached to the silica surface in two points (12.5% carbon content) giving a 1–12 pH stable stationary phase. The last column was a 150 mm × 3 mm I.D. Zorbax Sb-Aq also containing the ultra pure metal free 5 ␮m silica particle (180 m2 /g surface area, 8 nm pore diameter). The proprietary monofunctional diisopropyl silane bonded moiety was made hydrophilic by an embedded amide polar group designed to avoid any phase collapse in 100% aqueous mobile phase. The carbon content and bonding density were not given.

2.1. Chemicals 2.3. Apparatus and protocol The physico-chemical properties of the eight 1-alkyl-3methyl imidazolium salts that were used in this study were listed in Table 1 of the first article [26]. The eight ILs are noted R-MIM BF4 or R-MIM PF6 with the letters E, B, H and O standing respectively for 1-ethyl, butyl, hexyl and octyl-3-

The RPLC system was composed of a Shimadzu isocratic pump (model LC6A, Kyoto, Japan), a Shimadzu UV detector (model SPD6A) and an in-line Rheodyne 5025 valve with a 20 ␮l sample loop.

Table 1 Data for the log k vs. nC relationship obtained with the retention of alkyl-imidazolium salts in all studied systems log k = a nC + b

0.01 M Slopeb Interceptc r2 a b c

Kromasil C8

Zorbax Extend C18

50/50 acetonitrile–water

30/70 acetonitrile–water

NaCla

NaCla

0.175 −1.13 0.977

NaBF4 0.148 −0.71 0.992

NaPF6 0.152 −0.43 0.997

0.293 −1.77 0.966

NaBF4 0.305 −1.53 0.997

Zorbax Sb-Aq

NaPF6 0.319 −1.00 0.989

Regression parameters obtained with the alkyl-MIM BF4 retention factors only. The slope, a, is called the methylene selectivity for the 1-alkyl-3-methyl-imidazolium cations. The intercept, b, is the extrapolated log k for the methyl-imidazolium salt,

50/50

30/70 acetonitrile–water

NaPF6 0.164 −0.93 0.996

NaCla 0.208 −0.86 0.926

NaBF4 0.215 −0.83 0.992

50/50 NaPF6 0.221 −0.38 0.989

NaPF6 0.111 −0.44 0.997

478

M.J. Ruiz-Angel, A. Berthod / J. Chromatogr. A 1189 (2008) 476–482

The retention data were recorded using isocratic conditions with a flow-rate of 0.7 ml/min for the Kromasil C8 column and 0.4 ml/min for the Zorbax columns. These flow rates were producing equivalent mobile phase linear velocity taking in account the different internal diameters of the selected columns. Triplicate injections were made. The signal was monitored at 230 nm and recorded through an A/D converter by a personal computer running the Azur 4.6 Data Acquisition Sofware (Datalys, Grenoble, France). 3. Results and discussion 3.1. Column selection A recent work characterized HPLC packings with two physico-chemical properties: hydrophobicity of the packing and polar selectivity at neutral pH [27]. In this work, the hydrophobicity and polar selectivity index of the Kromasil C18 stationary phase were respectively, listed as 0.041 and −1.28. The conventional reversed-phase Kromasil C8 column was selected for its two different indexes: a lower hydrophobicity (0.0356) and

a higher polar selectivity (−1.23). The Zorbax Extend C18 was selected for its similar hydrophobicity index: 0.0405 associated to a very low polar selectivity: −1.50. The Zorbax Sb-Aq was selected for its very low hydrophobicity index: 0.0266, and polar selectivity: −1.45 [27]. 3.2. Stationary phase effects and hydrophobic interaction All the ILs analyzed exhibit reversed-phase behavior with the three selected stationary phases. Fig. 1 compares the chromatograms of the IL-PF6 salts separated on the three columns with the same mobile phase and similar linear flow velocity. The experimental log k values are linearly related to the acetonitrile volume percentage in the mobile phase [26,28–30]. Also, for a constant organic modifier concentration, the IL retention factors were related to nC , the carbon number of the imidazolium alkyl chain, by the following relationship for homologous series [31,32]: log k = a nC + b

(1)

where the slope, a, is the methylene selectivity and the intercept b is the log of the 1-methylimidazolium salt retention factor. Table 1 lists the slopes a, intercepts b and regression coefficients obtained on the different columns with the different mobile phases. In all cases, the regression coefficients, r2 , obtained for the k values of the ILs associated to the BF4 − anion were higher than those obtained for the PF6 − ILs when sodium chloride or hydrogen phosphate was used as added salt in the mobile phase. The methylene selectivities obtained with different added salts in the mobile phase but with the same stationary phase are very close (Table 1). There is a significant difference between the stationary phases; the methylene selectivity of the Kromasil C8 stationary phase is higher than the one of the two other stationary phases for the same mobile phase. In all cases (Fig. 1), distorted fronting or tailing peaks were observed meaning that linear chromatographic conditions are not fulfilled. In such column overload situations, the retention factors depend on injected concentrations. The obvious solution to improve peak shapes is to reduce the IL injected amounts to work on the linear part of the IL adsorption isotherms. However, distorted peaks and/or peak doubling were observed when the injected IL amounts were decreased especially with NaCl, NaHPO4 , NaBF4 and NaClO4 containing mobile phases. Further studies were needed to understand the chromatographic process. 3.3. Effect of the nature and concentration of added salts on ionic liquid retention factors

Fig. 1. Separation of hexafluorophosphate alkyl-3-methyl-imidazolium ionic liquids: EMIM, BMIM, HMIM (5 mM or 100 nmol injected) and OMIM (7 mM; 140 nmol); 20 (l injection volume; mobile phase: 50/50 (v/v) acetonitrile–water–0.01 M NaPF6 ; UV detection 230 nm. Columns and flow: (a) Kromasil C8 , 4 mm I.D.; 0.7 ml/min, (b) Zorbax Extend-C18 ; 3 mm I.D.; 0.4 ml/min and (c) Zorbax Sb-Aq; 3 mm I.D.; 0.4 ml/min.

3.3.1. Effect of the nature of the added salts Five different salts were studied as possible ionic additives to the acetonitrile/water mobile phases. Mobile phases containing 0.01 M of the salts: NaCl, NaH2 PO4 , NaBF4 , NaClO4 and NaPF6 , and 50/50 (v/v) acetonitrile–water were used with the Kromasil C8 column and 50/50 or 30/70 (v/v) acetonitrile–water for the Zorbax columns. No particular control of pH was made as

M.J. Ruiz-Angel, A. Berthod / J. Chromatogr. A 1189 (2008) 476–482

ILs do not show any acid–base behavior. Two mixtures of EMIM, BMIM, HMIM and OMIM associated to the anions BF4 − or PF6 − were injected separately. The injection of 20 ␮l of 3 mM of each ILs produced reproducible triangle shaped single peaks with the five different added salts. Table 2 compares the retention factors of the ILs (60 nmol injected) eluted with the selected mobile phases in the different systems With the Zorbax columns, the 50/50 (v/v) mobile phase did not produce enough retention for reliable k determination of the EMIM, BMIM and HMIM ILs. Except for the Na PF6 containing mobile phase, it was necessary to use a lower acetonitrile concentration of 30% (v/v) to observe some retention of the ILs. For all columns, the mobile phases containing the salt additives NaCl or NaH2 PO4 gave the smallest retention. The retention factors changed slightly in the presence of NaBF4 or NaClO4 and increased dramatically when NaPF6 was added to the mobile phase, especially for the long chain ILs. The Kromasil C8 column showed marked differences compared to the two Zorbax columns (Table 2).

479

observed. Fig. 2a shows the case of the Kromasil C8 column. The Zorbax Sb-Aq column produced similar results (not shown). The mobile phase ionic strength increase produces significant peak shape improvements with minimum changes in peak positions. The retention times (factors) of the OMIM PF6 most retained IL are 9.5 min (3.52), 9.7 min (3.62) and 10.3 min (3.9) with a 50/50 (v/v) acetonitrile/water mobile phase containing respectively 5, 10 and 15 mM NaPF6 added salt. Its peak widths at base are respectively, 4.2 min, 3.2 min and 2.8 min. The peak heights increase accordingly. Fig. 2b shows the same experiments done using the Zorbax Extend C18 column. The results are clearly different: the ionic strength increase produces a correlated increase in peak position with little changes in peak shape. For example, the retention times (factors) of the OMIM PF6 most retained IL are 3.8 min (2.17), 4.5 min (2.75) and 4.9 min (3.08) with respectively, 5, 10 and 15 mM NaPF6 added salt, the 2.4 min peak width at base and the peak heights remaining practically unchanged. 3.4. Retention mechanism

3.3.2. Effect of the concentration of the added salts The change of the concentration of the added salts produced a change in mobile phase ionic strength. Fig. 2 shows the chromatograms obtained when increasing concentrations of NaPF6 were added to the mobile phase. Two different effects were

These results must be associated with the results previously obtained [25,26] and results obtained by others [28–30]. The interaction of the alkyl-methylimidazolium cations with the alkyl bonded stationary phases occurs through a mixed mechanism involving hydrophobic and ionic interactions. Ionic liquids have a dual nature; they are trivially made by an IL anion and an IL cation. When an ionic liquid, noted [IL+, IL−], is injected in a mobile phase containing a buffer salt, noted [MP+, MP−], the four following salts can be found at the column head: [IL+, IL−] + [MP+, MP−] ↔ [IL+, IL−] + [IL+, MP−] +[MP+, IL−] + [MP+, MP−]

(2)

All four ions have different affinities for the stationary phase. However, none of the ions can move independently; electro neutrality should always be maintained meaning that a given cation must move with an anion and vice versa. Only the IL+ cation adsorbs UV light. Then only the two IL+ containing salts can be detected since non-UV absorbing salts are added to the mobile phase, and the IL− anion most often does not absorb UV light.

Fig. 2. Superimposed chromatograms of hexafluorophosphate alkyl-3-methylimidazolium ionic liquids (each 60 nmol injected) eluted with 50/50 (v/v) acetonitrile/water mobile phases containing 5 mM added NaPF6 (solid line); 10 mM added NaPF6 (dashed line); 15 mM added NaPF6 (dotted line); (a) Kromasil C8 , 4 mm I.D.; 0.7 ml/min, (b) Zorbax Extend C18 , 3 mm I.D.; 0.4 ml/min.

3.4.1. Ion adsorption The increase in retention factors for the alkyl-MIMs observed when changing the nature of the added salt, observed in Table 1 with the NaBF4 and especially NaPF6 containing mobile phases reflects the affinity of the inorganic anions for the stationary phases. It was found that the inorganic ions adsorb on the stationary phase in amount increasing according to the Hofmeister series which follows the order: Cl− ∼ H2 PO4 − < BF4 − ∼ ClO4 − < PF6 − [25]. Such inorganic anion adsorption seems to occur on all studied stationary phases. This anion adsorption is responsible for charge–charge interactions that are partly responsible for alkyl-MIM cation retention. These ionic interactions are stronger with the PF6 − anions widely adsorbed on the stationary phases but less intense with

480

M.J. Ruiz-Angel, A. Berthod / J. Chromatogr. A 1189 (2008) 476–482

Table 2 Retention factors of the investigated ionic liquids in the various chromatographic conditions studied (60 nmol injected, 10 mM added salt) k

Kromasil C8

Zorbax Extend C18

NaBF4 or NaClO4 b

NaPF6 b

NaPF6 a

0.3

0.4

0.4

1.3

0.6

0.7

0.3

0.3

0.5

1.2

0.6

0.4

1.4

0.5

0.6

0.9

2.8

1.0

0.9

1.0

1.8

0.5

1.5

1.6

2.7

1.0

2.9

0.6

2.1

8.1

1.1

1.1

2.8

8.2

1.7

1.5

2.9

2.1

2.2

7.7

1.1

2.7

2.9

8.3

1.7

2.0

3.2

6.4

5.0

8.3

39

2.6

5.0

8.2

27.5

3.0

2.1

3.2

6.4

5.2

7.9

38

2.5

5.2

8.0

27

2.9

NaCl or NaH2 PO4 a

NaBF4 or NaClO4 a

NaPF6

EMIM BF4 EMIM PF6 BMIM BF4 BMIM PF6 HMIM BF4 HMIM PF6 OMIM BF4 OMIM PF6

0.1

0.4

0.1

a b

Zorbax Sb-Aq NaCl or NaH2 PO4 b

0.01 M

a

NaCl or NaH2 PO4 b

NaBF4 or NaClO4 b

NaPF6

0.8

0.08

0.1

0.5

0.6

0.8

0.02

0.2

0.4

0.7

1.45

0.2

0.9

0.9

1.45

0.6

1.5

1.3

b

NaPF6

a

50/50 (v/v) acetonitrile–water with 0.01 M of the indicated salt. 30/70 (v/v) acetonitrile–water with 0.01 M of the indicated salt.

the weakly or non-adsorbed Cl− or H2 PO4 − anions resulting in lower retention times for the same organic modifier content (Table 2). 3.4.2. Adsorption isotherms The chromatographic mechanism is complex since it involves several adsorption isotherms. The first isotherm is the added anion adsorption isotherm: the amount of adsorbed ions increases with the concentration of the added salt in the mobile phase. This isotherm strongly depends on the stationary phase and the added salt [33,34]. Another adsorption isotherm is the imidazolium cations–anion covered stationary phase adsorption isotherm. The tailing peak shapes observed in the Figs. 1 and 2 chromatograms suggest a convex shape for these IL-modified stationary phase isotherms. The two isotherms are not independent. The increases in retention times observed in Fig. 2 are due to the larger amount of adsorbed anions on the stationary phases (first isotherm). The changes in peak shapes are associated with changes in the shape of the second adsorption isotherm. The curvature of the Kromasil and Zorbax Sb-Aq combined adsorption isotherms seem to decrease when the PF6 − concentration increases (better peak shapes in Fig. 2a). There is no such change in the same conditions with the Zorbax Extend C18 stationary phase (no change in peak shapes in Fig. 2b). 3.4.3. Ion-exchange and ion-pairing An imidazolium cation is necessarily associated with an anion. Its retention can be due to hydrophobic retention of the ion-pair as well as ion-exchange with the anion covered stationary phase surface. These two mechanisms were experimentally observed when working with the less chaotropic added salts NaCl and NaH2 PO4 . With mobile phases containing 0.01 M NaCl or 0.01 M NaH2 PO4 surprising differences in the chromatograms, in the form of distorted peaks, were observed in

the analyses of ILs differing by the anions BF4 − or PF6 − . This behavior was fully studied with the 0.01 M NaCl mobile phases and all three tested stationary phases. The experimental chromatograms are shown in Fig. 3 for the Kromasil column and the mobile phase 50/50 (v/v) acetonitrile–water–0.01 M NaCl. The

Fig. 3. Superimposed chromatograms of increasing amounts of: (a) EMIM, BMIM, HMIM and OMIM BF4 and (b) EMIM, BMIM, HMIM and OMIM PF6 on a Kromasil C8 column. Concentration range: 1–11 mM in 2 mM steps (20–220 nmol injected in 40 nmol steps) except for OMIM, which was 3–13 mM (60–260 nmol injected); mobile phase 50/50 (v/v) acetonitrile–water–0.01 M NaCl; 0.7 ml/min, UV detection 230 nm.

M.J. Ruiz-Angel, A. Berthod / J. Chromatogr. A 1189 (2008) 476–482

chromatograms obtained when increasing amounts of the ILs in the tetrafluoroborate form (Fig. 3a) or the hexafluorophosphate form (Fig. 3b) were injected on the Kromasil C8 column are overlaid with a common time axis. Very similar profiles were obtained for the Zorbax columns and/or when 0.01 M NaH2 PO4 was used as the salt additive.. The reason proposed for this behavior is the very weak adsorption of the Cl− or H2 PO4 − anions on the stationary phases compared to the BF4 − and especially PF6 − anion adsorption [25]. When the mixture of tetrafluoroborate (Fig. 3a) or hexafluorophosphate (Fig. 3b) ILs is injected in a chloride containing mobile phase, the chaotropic anions BF4 − or PF6 − are somewhat retained due to their affinity for the stationary phase [26]. These anions must be associated with cations to maintain electroneutrality. In Eq. (2), the mobile phase anion, MP− , is the chloride anion. As long as its concentration (0.01 M) is higher than the concentration of the chaotropic injected anions, either BF4 − or PF6 − , a total exchange of ions is possible: the chaotropic injected anions elute as undetected sodium salts and the imidazolium cations elute as UV detected chloride ionpairs. When the concentrations of the injected ILs increase, the local injected anion concentration may pass the 0.01 M mobile phase sodium chloride concentration. Then, there are not enough sodium cations to associate with all the chaotropic injected anions and parts of them stay electrically associated with IL+ cations. Since the IL+ Cl− ion-pair move faster than the IL+ PF6 − ion-pair, the IL+ peak is seen broadening and even splitting in two. The peak splitting effect is clearly seen with the EMIM BF4 salt (Fig. 3a). At low injected concentrations, the EMIM cations show up at a retention time of 1.43 min (k = 0.1) corresponding to the EMIM Cl salt. As the injected concentrations increased, the peak at 1.43 min stops to increase and a second peak at 1.8 min (k = 0.38) develops. It corresponds to the EMIM BF4 salt. This phenomenon does not occur with the three other IL+ cations due to the chromatographic process. These three cations are more retained by the stationary phase than the BF4 − anion and there is a continuous input of fresh mobile phase with chloride anions. The three cations move in the column as chloride salts with different adsorption isotherms. The BMIM Cl IL shows a constant retention time at 1.9 min (k = 0.46) and almost symmetric shape (almost linear isotherm). The HMIM Cl peak elutes also at an almost constant retention time (2.25 min, k = 0.73) with a somewhat tailing shape (slightly convex isotherm). The OMIM Cl peaks show an increasing retention time with a clearly fronting shape characteristic of a strongly concave adsorption isotherm [26,33,34]. The chaotropic character of the PF6 − anion is significantly higher than that of the BF4 − anion [25]. It means that the former anions are more retained by the hydrophobic C8 stationary phase than the later ones. Fig. 3b shows that increasing amounts of PF6 − injected anions interfere with the two BMIM+ and HMIM+ cation elution so that both show the peak splitting effect described for the EMIM+ cations with BF4 − injected salts. This argument permits to identify the emerging peaks at 2.4 and 3.0 min in Fig. 3b as BMIM and HMIM PF6 . The EMIM+ cations elute with chloride anions more rapidly than the PF6 −

481

anions in a single peak increasing regularly with the injected amount. Conversely, the OMIM+ cations are more retained than the PF6 − anions and also elute in a single peak with chloride anions coming from the continuous flow of mobile phase. Fig. 3a and b show that the EMIM and OMIM ILs both elute at the same retention time whatever ionic form is initially injected. 3.5. Quantitative analysis Quantitative IL analyses and retention mechanism are intimately linked. As previously found by Stepknowski et al. [24,28] and confirmed Ruiz-Angel and co-workers [25,26], ILs differing by the cation were clearly separated in RPLC when an added salt is present in the mobile phase. The lowest limits of detection (LODs) were obtained with the Kromacil C8 column and the NaPF6 added salts. With this column and the acetonitrile–water–0.01 M NaPF6 50/50 (v/v) mobile phase, the LODs were respectively 1.6, 2, 3 and 5 nmol injected for the EMIM, BMIM, HMIN and OMIM ionic liquids. These LODs respectively correspond to 180 ng, 280 ng, 0.5 ␮g and 1.2 ␮g injected mass of the cations. These LODs are one order of magnitude lower than the LODs obtained by ionexchange chromatography [28]. Comparable but slightly higher LODs were obtained with the same mobile phase and the two Zorbax columns. When NaClO4 or NaBF4 were used as additives, slight differences in the retention factors of some ILs depending on the associated anion in the injected sample (e.g. BMIM Cl differed somewhat from BMIM PF6 , Table 2). With all columns, the obtained peaks were always broader than those obtained with the mobile phases containing NaPF6 added salt. Consequently the LODs were poorer that those previously obtained with the NaPF6 additive. 4. Conclusion From a quantitative point of view, the use of additives with significant affinity for the stationary phase, i.e. salts with chaotropic anions such as NaPF6 , are strongly recommended for the analysis of ILs by RPLC with UV detection. These chaotropic added salts totally minimize the differences among ILs associated to different anions and allows for a significant retention of the low molecular weight and polar IL members. Hexafluorophosphate anions could slowly hydrolyze releasing fluoride anions. However, no particular column damage or irreversible adsorption was observed by the use of NaPF6 as additive after several months of work with the different columns presented in this work. The usual column care procedure: rinse with salt-free mobile phase every night and storage in pure organic solvent over the week-end, was followed. Nevertheless, this salt may be substituted by others with acceptable affinity for the stationary phase (e.g. NaBF4 or NaClO4 ). Perchlorate anions are good chaotropic anion that could be associated with ammonium cations if mass spectrometry must be the detection tool. It recalled that MS is certainly not the recommended detector when analyzing ILs. Indeed, ILs have an extremely low vapor pressure and could definitely condense in the MS ionization

482

M.J. Ruiz-Angel, A. Berthod / J. Chromatogr. A 1189 (2008) 476–482

source polluting it. Kosmotropic salts such as NaCl or NaH2 PO4 are unadvisable for RPLC IL analysis even if phosphate anions are needed for buffering reasons because the obtained chromatograms will depend on the IL anions and on the injected amounts. Considering a possible stationary phase effect, the results offered by the conventional Kromasil C8 column were very similar to those obtained with the Zorbax columns in terms of selectivity, resolution and peak shape. The ionic interactions with the two IL constituents, anion and cation, within the mobile phase and the ionic interactions with the stationary phase seem to screen the differences in hydrophobicity and polar character between the three selected stationary phases. Compared to the Kromasil column, the Zorbax columns permit to reduce the analysis duration, but the selection of special bidentate end-capped columns (Zorbax Extend-C18 ) or columns with polar packings (Zorbax Sb-Aq) seems not to be especially necessary for the chromatographic analysis of ILs when appropriate chaotropic salts are added at a concentration high enough in the mobile phase. Acknowledgements M.J.R.A. is grateful to Ministerio de Educaci´on y Ciencia from Spain for a post-doctoral fellowship, a Ram´on y Cajal position and also for financial support by Project CTQ200402760/BQU. A.B. thanks the Centre National de la Recherche Scientifique (CNRS, UMR 5180 – Pierre Lanteri), for continuous support. References [1] R.D. Rogers, K.R. Seddon, Ionic Liquids, ACS Symposium Series, American Chemical Society, Washington, DC, vol. 818, 2002; vol. 856, 2003; vol. 902, 2005. [2] C.F. Poole, J. Chromatogr. A 1037 (2004) 49. [3] J.G. Huddleston, A.E. Visser, W.M. Reichert, H.D. Willauer, G.A. Broker, R.D. Rogers, Green Chem. 4 (2001) 156. [4] K.R. Seddon, J. Chem. Technol. Biotechnol. 68 (1997) 351.

[5] A. Kuwar, Chem. Rev. 101 (2001) 1. [6] S. Carda-Broch, A. Berthod, D.W. Armstrong, Anal. Bioanal. Chem. 375 (2003) 191. [7] D.W. Armstrong, L.F. He, Y.S. Liu, Anal. Chem. 71 (1999) 3873. [8] A. Berthod, L.F. He, D.W. Armstrong, Chromatographia 53 (2001) 63. [9] S. Carda-Broch, A. Berthod, D.W. Armstrong, Rapid Commun. Mass Spectrom. 17 (2003) 553. [10] E.G. Yanes, S.R. Gratz, M.J. Baldwin, S.E. Robinson, A.M. Stalcup, Anal. Chem. 73 (2001) 3838. [11] T.N. Laremore, S. Murugesan, T.J. Park, Y.F. Avci, D.V. Zagorevski, R.J. Lindhart, Anal. Chem. 78 (2006) 1774. [12] T.F. Jiang, Y.L. Gu, B. Liang, J.B. Li, Y.P. Shi, Q.Y. Ou, Anal. Chim. Acta 479 (2003) 249. [13] L. He, W. Zhang, L. Zhao, X. Liu, S. Jiang, J. Chromatogr. A 1007 (2003) 39. [14] X. Xiaoua, Z. Liang, L. Xia, J. Shengxiang, Anal. Chim. Acta 519 (2004) 207. [15] M.J. Ruiz-Angel, S. Carda-Broch, A. Berthod, J. Chromatogr. A 1119 (2006) 202. [16] Y. Franc¸ois, A. Varenne, E. Juillerat, D. Villemin, P. Gareil, J. Chromatogr. A 1155 (2007) 134. [17] C.J. Tran, D. Oliveira, S.F. Yu, Anal. Chem. 78 (2006) 1349. [18] M.J. Ruiz-Angel, V. Pino, S. Carda-Broch, A. Berthod, J. Chromatogr. A 1151 (2007) 65. [19] A. Berthod, S. Carda-Broch, Anal. Bioanal. Chem. 380 (2004) 168. [20] J.L. Anderson, D.W. Armstrong, G.T. Wei, Anal. Chem. 78 (2006) 2892. [21] S. Pandey, Anal. Chim. Acta 556 (2006) 38. [22] J.F. Liu, J.A. Jonsson, G.B. Liang, TrAC Trends Anal. Chem. 24 (2005) 20. [23] W. Qin, H. Wei, S.F.Y. Li, The Analyst 127 (2002) 490. [24] P. Stepknowski, A. M¨uller, P. Behrend, J. Ranke, J. Hoffmann, B. Jastorff, J. Chromatogr. A 993 (2003) 173. [25] A. Berthod, M.J. Ruiz-Angel, S. Huguet, Anal. Chem. 77 (2005)4071. [26] M.J. Ruiz-Angel, A. Berthod, J. Chromatogr. A 1113 (2006) 101. [27] U.D. Neue, K.V. Tran, P.C. Iraneta, B.A. Alden, J. Sep. Sci. 26 (2003) 174. [28] P. Stepknowski, W. Mrozik, J. Sep. Sci. 28 (2005) 149. [29] S. Kowalska, B. Buszewski, J. Sep. Sci. 29 (2006) 2625. [30] B. Buszewski, S. Kowalska, P. Stepnowski, J. Sep. Sci. 29 (2006) 1116. [31] H. Engelhart, M. Jungheim, Chromatographia 29 (1990) 59. [32] E. Cruz, M.R. Euerby, C.M. Johnson, C.A. Hackett, Chromatographia 44 (1997) 151. [33] F. Gritti, G. Giochon, Anal. Chem. 76 (2004) 4779. [34] F. Gritti, G. Giochon, J. Chromatogr. A 1028 (2004) 197.