Ionic liquids (ILs) are organic molten salts formed by bulky dissymmetrical organic cations and various anions that are liquids at ambient temperature. They show specific physicochemical properties that make them more versatile and useful to applications in various areas of chemical fields. For example, they can be used as solvents in organic synthesis and catalytic processes[1, 2], enzyme-catalyzed reactions[3, 4], electrochemical studies [5, 6], and liquid-liquid extractions [7, 8]; they can also be used as liquid matrices for matrix-assisted laser desorption/ionization mass spectrometry [9, 10] and running electrolytes [11, 12] in addition to serving as electrolyte additives in capillary electrophoresis(CE) [13, 14], mobile phase additives in liquid chromatography [15, 16], and stationary phases in gas chromatography (GC) and high-performance liquid chromatography (HPLC) [17-20].
ILs exhibit desirable properties including high thermal stability, high viscosity, non-volatility, and non-flammability. Therefore, they are great choices as GC stationary phases. The first application of ILs as a GC stationary phase was reported by Barber et al. [21]. Then, alkylammonium-and alkylphosphonium-based stationary phases were prepared and also investigated [22, 23]. However, these phases have some inherent limitations, such as thermal instability, relatively narrow liquid ranges, and poor wettability toward the surface of fused silica. Alkylimidazolium-and alkylpyridinium-based ILs have also been used as GC stationary phases [17, 18]. Initial studies by Armstrong et al.[17] indicated that room temperature ionic liquids (RTILs) possess "dual nature" properties. They could be used as non-polar or polar stationary phases for separation of non-polar or polar compounds, respectively.
However, there is very little research reported on the use of ILs as HPLC stationary phases. The first application of ILs as a stationary phase for HPLC was reported by Liu et al. in 2004 [24]. The silica-based vinyl hexylimidazolium tetrafluorobarate stationary phase was prepared. It was successfully applied to the separation of ephedrines. Then, Stalcup and coworkers [25] studied the retention characteristics of a butylimidazolium-based stationary phase through linear solvation free energy relationships. They found that the retention characteristics of the test solutes on the new stationary phase indicated significant similarities to phenyl stationary phases despite the presence of a positive charge on the imidazolium phase. Two different alkylimidazoles were attached to a silica support, and other IL-based stationary phases were synthesized by Colón and coworkers [20]. These phases were used for the separation of aromatic carboxylic acids. Therefore, retention properties of the stationary phases were evaluated. It was observed that the separation mechanisms involved multiple interactions.
A group of approximately 20 surface-confined IL (SCIL) stationary phases were prepared and evaluated recently [26]. Mostly, alkylimidazolium-and pyridinium-based ILs were used to chemically modify the surface of silica particles to act as the HPLC stationary phase. In our previous work [27], N-octylimidazolium bromide and N-octylimidazolium tetrafluoroborate ILs immobilized on silica were tested as HPLC stationary phases.
In the present study, a derivative of N-pyrrolidinium was covalently immobilized on the surface of silica particles to prepare a new stationary phase, for the first time to our know-ledge. In fact, we synthesized a novel stationary phase based on N-methyl pyrrolidinium tetrafluoroborate attached on the silica surface. Then, the modified particles were characterized and tested under reversed-phase chromatographic conditions. The new stationary phase was also compared with our previously reported phase. It was found that this novel stationary phase exhibited strong ion-exchange character and coexisting interactions were observed in reversed-phase HPLC.
The 5-μm Lichrosorb Si 100 porous silica, with an average specific surface area of 300m2/g, was purchased from Merck and used as support.(3-Bromopropyl)trimethoxysilane was obtained from Sigma-Aldrich (MO, USA). N-Methyl pyrrolidine was synthesized in Chemistry and Chemical Engineering Research Center of Iran (CCERCI, Tehran, Iran). All HPLC-grade solvents were purchased from Merck (Darmstadt, Germany). They were filtered with a 0.45 μm membrane filter before injection into the HPLC system. The reaction solvents, including p-xylene and diethylether, were dried over sodium-potassium and distilled before use. The model compounds, including benzene, alkyl benzenes, anthracene, naphthalene, N, N-dimethyl aniline, p-nitro aniline, phenol, chlorobenzene, iodobenzene, 4-hydroxy benzoic acid, and benzoic acid, were analytical grade and obtained from Merck (Darmstadt, Germany). All of these compounds were dissolved in solutions with the same mobile phase composition.
The synthesis procedure is the same as that described in our previous paper [27]. First, silica was dried at 140 ℃ for 24 h and cooled in a desiccator to remove adsorbed water. Then, 4.25 g of activated silica particles under an Ar atmosphere were suspended in 50 mL of dry p-xylene. An excess of (3-bromopropyl)trimethoxysilane(2.5 mL) was added. The suspension, under inert gas, refluxed for 48 h, while stirring was done with bubbling Ar. After refluxing, the modified silica was collected by filtration and exhaustively washed with dry diethylether. Next, the modified silica was washed with a mixture of methanol/water (50:50, v/v) and suspended in water for 90 h to promote the total hydrolysis of the remaining methoxy groups. The obtained bromopropyl silica (SilprBr) was then endcapped with excess chlorotrimethylsilane (4 mL) in dry p-xylene (60.0 mL). After refluxing for 24 h, the reaction was stopped, and the endcapped silica (EsilprBr) was filtered by vacuum glass filter under Ar. It was then washed with dry diethylether and tetrahydrofuran and dried for 12 h at 120 ℃.
Second, the chemically bonded bromopropyl group on the silica surface reacted with N-methyl pyrrolidine (1 g) in dry p-xylene (86 g) as solvent. The mixture under Ar inert atmosphere was refluxed for 40 h. After refluxing, the reaction was stopped and the modified silica was cooled to room temperature. Then, it was transferred to a vacuum glass filter and washed with methanol and tetrahydrofuran and dried for 12 h at 100 ℃ to yield the silica chemically bonded with N-methyl pyrroldinium Br (SilprMP Br).
Third, the Br- anion was converted to BF4- by adding 1 g of NaBF4 to SilprMP Br and shaking for 24 h in 50 mL of dichloromethane. Then, the suspension was filtered and washed with diethyl ether, methanol, and water and dried at 100 ℃ for 24 h; thus, SilprMP BF4 was obtained.
Elemental analyses were performed on a Flash EA 1112 CHNS from Thermo Finnigan-CE instrument (Milan, Italy) for the synthesized materials. At least two determinations were made for each compound.
Thermogravimetric analyses were carried out by a Netzsch TG 209 F1 equipment (Selb, Germany). All thermogravimetric measurements were performed under a N2 atmosphere over the temperature range of 20 to 900 ℃ with a heating rate of 10 ℃/min. At least two determinations were carried out for each material.
Diffuse reflectance infrared Fourier transformation spectra of modified particles were obtained on a Bruker IFS 48 FTIR spectrometer (Ettlingen, Germany) in the range of 4 000-1 400 cm-1.
Stainless-steel column (125 mm×4.6 mm) was downward packed using the stationary phase's slurries in isopropanol-tetrahydrofuran. The methanol was applied as the propulsive solvent. A packing pressure of 55 MPa (by Knauer Well Chrom K-1900 Pneumatic Pump) was used. Then, the packed column was conditioned by passing 50 column volumes of methanol, to prepare the column to be use.
All chromatographic tests were performed on a Knauer Well Chrom including a K-1001 Pump and a K-2600 UV-Vis detector (Berlin, Germany) at room temperature. The flow rate was 1 mL/min. The UV wavelength was set at 254 nm. The system was equipped with a model 7125 Rheodyne manual injector (Coronado Island, USA) and also with a 20 μL sample loop. Methanol, acetonitrile, deionized water, and phosphate buffer (100 mmol/L) at different pH values ranging from 3.2 to 6.8 were used as the mobile phases. Buffer pH was adjusted by phosphoric acid (100 mmol/L) and sodium hydroxide (100 mmol/L). The dead time of the column was determined by the mobile phase signal in the UV detection.
Retention characteristics of the stationary phase were studied by using a group of compounds that included aromatics, alkyl benzenes, phenol, halogenated benzenes, and acidic and basic compounds. Chromatographic parameters, including retention factor (k), separation factor (α), and plate number (N), were calculated using the separation of the mentioned compounds.
A schematic diagram of the preparation of SilprMP BF4 is shown in Fig. 1. First, silica was silanized with (3-bromopropyl)trimethoxysilane, under inert gas, to yield bromopropyl silica (SilprBr). The SilprBr was washed with water to promote total hydrolysis of the remaining methoxy groups and was then endcapped. A large excess of N-methyl pyrrolidine was reacted with bromopropyl groups to produce N-methyl pyrroldinium Br. The role of the solvent is important from the reaction yield point-of-view. p-Xylene (boiling point=138.4 ℃) was suitable as a reflux solvent, which led to better efficiency of the synthesis reaction and greater yield compared with other used solvents that were reported in literature, such as toluene [26].
Finally, the phase anion was changed to BF4- by reaction of SilprMP Br with NaBF4, and SilprMP BF4 was obtained. Anion replacement from Br to BF4 improved the separation of non-polar compounds but reduced the separation factor of polar compounds [27]. This study also led to similar results; BF4 was observed to be a softer anion, leading to better resolution for non-polar and/or low polar compounds.
Elemental analysis, Fourier transform infrared spectrometry, and thermogravimetric analysis were used for characterization of the prepared materials, resulting from the availability of facilities.
The elemental analysis data can be employed to prove immobilization on the silica surface. It can be used to estimate the surface coverage of the chemically modified silica as well. The carbon, hydrogen, and nitrogen elemental analysis results and the surface coverage for SilprBr, EsilprBr, and SilprMP Br4 are summarized in Table 1. As can be seen, after silica modification, the carbon, hydrogen, and nitrogen percentages of the functionalized silica all clearly increased. In equations (1) and (2),
M, m(C), m(H), m(N), and S represent the molecular weight of the bonded molecule, the carbon, hydrogen, and nitrogen mass percentages (%), and the specific surface area of silica support (300 m2/g), respectively [28, 29]. NN and NC also demonstrate the number of nitrogen and carbon organic groups attached to the silica surface. The mass percentage of carbon was employed in equation (1), and the concentration of the organic groups attached to the silica surface was calculated as 2.22 and 3.94 μmol/m for SilprBr and EsilprBr, respectively. Also, based on the mass percentage of nitrogen and applying equation (2), the bonding density was calculated as 1.14 μmol/m for SilprMP BF4.
TGA is used to determine the thermal stability of SCIL phases. Also, it can be used to confirm immobilization of the organic groups attached to the silica surface. Fig. 2 shows the thermogravi-metric curves for bare silica, EsilprBr, and SilprPM BF4. Three distinct weight losses between 180 ℃ and 900 ℃ were found to be associated with the loss of the organic groups attached to the surface. The initial weight loss at a temperature below 100 ℃ for all samples was attributed to the removal of physically adsorbed water and/or any minor solvent residues remaining from the synthesis or washing procedures. Physically adsorbed water was removed completely through further heating at approximately 180 ℃. The second loss of weight occurred in the range of 180-420 ℃. The weight loss observed in this range was attributed to the loss of the organic groups attached to the surface [30]. Hence, SilprPM BF4 displayed a higher mass loss than silica and EsilprBr, according to N-methyl pyrroldinium tetrafluoro borate binding. Other losses observed in the range of 420-900 ℃ were associated with decomposition of the residual methoxy side groups [31].
Infrared spectroscopy is known to be a useful tool for the identification of chemical modifications. In the spectrum of this phase, an important band at 1 565 cm-1 was attributed to the characte-ristic wave number of the C-N bond, which confirmed the anchoring of the organic molecule onto the silica surface.
First, a 125 mm×4.6 mm HPLC column was packed with the SilprMP BF4 modified silica particles, and a large number of separations was performed under reversed-phase conditions to evaluate the potential benefits and retention properties of the stationary phase.
Stationary phase hydrophobicity is measured by methylene group selectivity. This reflects the ability of the phase to separate two molecules that differ only in methylene group (e. g., homologues of alkylbenzenes). Therefore, the first investigation was performed with a test mixture composed of ethyl benzene, butyl benzene, pentyl benzene, and hexyl benzene. Successful separation of alkyl benzene compounds in the sample mixture was obtained using methanol/water as the mobile phase under gradient conditions as can be seen in Fig. 3.The retention of alkyl benzenes increased with the increasing number of carbons in the long alkyl chain, which indicates that hydrophobic interaction played the prominent role in the separation. The same results were obtained when a mixed solution containing benzene, toluene, and ethyl benzene was prepared and separated on the SilprMP BF4 stationary phase using water/methanol as the mobile phase. According to the observations, it can be concluded that this interaction in the stationary phase was effective enough to separate the methylene group. This reflects the importance of hydrophobic interaction in SilprMP BF4.
A second chromatographic evaluation was performed using a mixture composed of benzene, naphthalene, and anthracene on a SilprMP BF4 column. Fig. 4 shows a chromatogram of this test mixture using water/acetonitrile (80:20, v/v) as the mobile phase. From the chromatogram, it can be seen that, the slowest and fastest retention time were obtained for anthracene and benzene, respectively. This elution order can be explained by the cation-pi interaction of the pyrrolidinium cation of SilprMP BF4 with the aromatic ring(s) of the solute. Anthracene contains three rings and therefore has more interaction with the stationary phase; consequently, it eluted later than naphthalene and benzene, which have less interaction with the stationary phase and spend more time in the mobile phase. Also, the hydrophobic interaction of the stationary phase may help to get this elution order. Both interactions play significant roles in this separation.
Additionally, varying the amounts of water in the mobile phase and the subsequent effect on the retention of aromatic and alkyl benzene compounds was studied. We observed that the retention times of the solutes were associated with the water content in the eluent. Retention times increased as the amount of water in the mobile phase increased. Inversely, the retention time of the compounds decreased with an increase in the organic content of the mobile phase. These features reflect the hydrophobic nature of the new stationary phase.
The separation capability of some benzene derivatives was examined as well. Phenol, chlorobenzene and iodobenzene were separated successfully on this column. The retention and separation factors of the test mixture on SilprMP BF4 are listed in Table 2.
It can be seen that the retention factor of iodobenzene (2.27) is much larger than that of phenol (0.29) and chlorobenzene (1.17). However, iodobenzene is more hydrophobic than chlorobenzene and phenol, but hydrophobicity differences between analytes are not adequate to explain this observed separation by hydrophobic interaction. Therefore, we suggested that the ion-dipole interaction between the cation of the stationary phase and dipolar compounds was also responsible. This interaction could bring about positive effects in this separation. Iodobenzene has the greatest dipole moment (5.336×10-30 ℃·m), while this factor for chlorobenzene and phenol are 5.136×10-30 and 4.069×10-30 ℃·m, respectively. Consequently, the less polar compound will spend more time in the solvent, and will therefore elute first.
Altogether, we concluded that the hydrophobic interaction was the dominant interaction, and the ion-dipole interaction was also an effective interaction mechanism with the stationary phase for separation of molecular compounds.
To investigate the ionic interaction of the new stationary phase, acids and bases were separated on the column using a range of phosphate buffer without any organic solvent as mobile phases. The pH values of the mobile phases were also varied from 3.1 to 6.4. pH values lower than 3.0 were avoided to prevent the hydrolysis of siloxane bonds and consequently loss of the bonded phase.
The separation of a test mixture of aromatic acids composed of 4-hydroxy benzoic acid (pKa=4.6) and benzoic acid (pKa=4.2) using phosphate buffer at pH 3.1 was carried out (pKa=-log (acid costant)). The elution order corresponded to pKa. The solute retention factors increased with decreasing pKa values, which indicated a strong dependence on the ionic character of the test solutes. At this pH, near the pKa of benzoic acid, the extent of molecular and ionic forms of solute was approximately equal. So, both anion-exchange and molecular interactions were effective. In contrast, 4-hydroxy benzoic acid is less ionized. Hence, the ionic retention of solute with the cationic moiety of the stationary phase was reduced.
Then, the stationary phase was evaluated in detail to determine the effects of mobile phase pH on retention factors of the test mixture (increasing pH from 3.1 to 6.4). The effect of pH changes on retention factors can be seen in Fig. 5. The retention factors of the acidic test solutes decreased with increasing pH, which was attributed to the decreasing molecular state of compounds. Thus, pH changes reduced hydrophobic and ion-dipole interactions. In fact, at pH values lower than the pKa of the test compound, the retention increased owing to participation of stronger molecular interactions in the retention process as the main retention mechanism. But, at pH values greater than the solute's pKa, the ionization of the test compounds increased and molecular interactions were minimized. These results indicated that molecular interactions rather than ionic interactions were stronger in SilprMP BF4. The chromatographic parameters of this separation at pH values of 3.1, 5.3, and 6.4 were calculated and summarized in Table 3.
An investigation into the effects of varying pH on the retention factors of organic bases was performed as well to determine the extent of cation-exchange and electrostatic interactions with the SilprMP BF4 stationary phase. Therefore, the test mixture consisting of N, N-dimethyl aniline (pKa=5.1) and p-nitro aniline (pKa=1) was separated on the new phase using phosphate buffer with varied pH values from 3.1 to 6.4 as the mobile phase.
At pH 6.4, two organic bases were separated with good resolution and low asymmetries (Fig. 6). At this pH, the basic compounds were much less ionized, hence ionic and electrostatic interactions of the stationary phase did not affect their retention; molecular interactions played a major role in this separation. The retention time was longer for N, N-dimethyl aniline, which is more polar, while p-nitro aniline eluted more readily owing to less dipole interaction with the phase.
During more testing, separation of these analytes was performed at low pH values. At pH=5.2, N, N-dimethyl aniline has a partially positive charge that will lead to electrostatic repulsion between ionized specimens and pyrrolidinium cations. Under such conditions, the retention of N, N-dimethyl aniline decreased considerably, and the elution order was inverted. As the pH was reduced further to 3.2, N, N-dimethyl aniline had a full positive charge that resulted in increased electrostatic interactions with the stationary phase, thus reducing retention. For p-nitro aniline, its ionization did not change significantly over the studied pH range. Therefore, the retention time remained essentially unchanged. According to experimental results, the new stationary phase was not able to undergo cation exchange. The effect of the different pH values on the retention factor of basic test solutes on the new stationary phase is represented in Table 4.
Finally, it can be emphasized that separation of acidic and basic compounds was successful using buffer solution without organic solvent, which may be advantageous for the new stationary phase.
N-Methyl pyrrolidinium was bonded to silica particles to prepare a novel HPLC stationary phase. Thermogravimetric and elemental analyses corroborated the successful attachment of the derivatives to the silica particles. Using this new phase, aromatics, alkyl benzenes, and acidic and basic compounds (in addition to some derivatives of benzene) were separated successfully. Desirable column efficiency and satisfactory resolution were obtained. Thus, in comparison with previous stationary phases mentioned in Heydar et al. [27], the new stationary phase demonstrated better properties (e. g., ability to separate short alkyl chains). In addition, an improvement in separation factor was observed. The retention mechanism of this stationary phase and the effect of eluent pH were investigated. The separation mechanism appears to involve multiple interactions, including anion exchange and hydrophobic and electrostatic interactions, while hydrophobic interaction plays a significant role despite the positive charge on the pyrrolidinium cation. This multimodal retention behavioral ability to interact with analytes through different mechanisms is similar to previously reported SCIL phases. Potentially, the new stationary phase can be applied to separate many compounds using a small volume of organic solvent so that the acidic and basic solutes are separated using only with a single buffer solution.
2017,
Vol. 35 Issue (11): 1120-1128
PDF
2017,
Vol. 35 Issue (11): 1120-1128
PDF