Journal of Chromatography A, 1342 (2014) 54–62
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Solvent systems with n-hexane and/or cyclohexane in countercurrent chromatography—Physico-chemical parameters and their impact on the separation of alkyl hydroxybenzoates Michael Englert, Walter Vetter ∗ University of Hohenheim, Institute of Food Chemistry, Garbenstrasse 28, D-70599 Stuttgart, Germany
a r t i c l e
i n f o
Article history: Received 13 January 2014 Received in revised form 28 February 2014 Accepted 19 March 2014 Available online 27 March 2014 Keywords: Countercurrent chromatography Solvents Separation efficiency Alkyl hydroxybenzoates Physico-chemical parameters
a b s t r a c t Countercurrent chromatography (CCC) is an efficient preparative separation technique based on the liquid–liquid distribution of compounds between two phases of a biphasic liquid system. The crucial parameter for the successful application is the selection of the solvent system. Especially for nonpolar analytes the selection options are limited. On the search for a suitable solvent system for the separation of an alkyl hydroxybenzoate homologous series, we noted that the substitution of cyclohexane with n-hexane was accompanied with unexpected differences in partitioning coefficients of the individual analytes. In this study, we investigated the influence of the subsequent substitution of n-hexane with cyclohexane in the n-hexane/cyclohexane/tert-butylmethylether/methanol/water solvent system family. Exact phase compositions and polarity, viscosity and density differences were determined to characterize the different mixtures containing n-hexane and/or cyclohexane. Findings were confirmed by performing CCC separations with different mixtures, which led to baseline resolution for positional isomers when increasing the amount of cyclohexane while the resolution between two pairs of structural isomers decreased. With the new methodology described, structurally similar compounds could be resolved by choosing a certain ratio of n-hexane to cyclohexane. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Countercurrent chromatography (CCC) is a versatile and efficient preparative separation technique for the purification of various natural [1–5] and synthetic products [6–8] from crude mixtures as well as for the isolation of pharmaceutical drugs [9,10]. The chromatographic separation is based on the liquid–liquid distribution of analytes between mobile and stationary phases which represent two immiscible liquids [11,12]. Centrifugal fields are used to maintain the stationary phase in the column while the mobile phase is pumped through it to elute the analytes according to their partitioning coefficients (K) [13]. The selection of an appropriate solvent system is of utmost importance for the successful application and comparable to the role of both, the column and an eluent in high-performance liquid chromatography (HPLC) [12,14,15]. Numerous examples of twophase solvent systems are described that have been successfully employed for separations over the years in the literature [14,15].
∗ Corresponding author. Tel.: +49 711 459 24016; fax: +49 711 459 24377. E-mail address: [email protected] (W. Vetter). http://dx.doi.org/10.1016/j.chroma.2014.03.050 0021-9673/© 2014 Elsevier B.V. All rights reserved.
In general, three important criteria should be fulfilled by the solvent system for successful CCC separations. First, it should form two clear distinct phases with a separation time (ts ) less than 30 s to preclude emulsification [11,16]. Second, the selected solvent system should provide a high retention of the stationary phase, measured by the stationary phase retention (Sf ) [11]. Third, the sample compounds should have partitioning coefficients K between 0.4 and 2.5 with significant differences between the individual analytes, corresponding to high chromatographic selectivity [13,17]. As reported in a recent review, only a restricted number of solvents are employed in CCC with water, n-hexane, ethyl acetate and methanol as the top 4 utilitarian solvents [18]. Especially for nonpolar analytes the choice of solvent mixtures is very limited. The less polar solvent is usually represented by an alkane. Berthod et al. demonstrated that different alkanes (n-pentane, n-hexane, cyclohexane and iso-octane) used in place of n-heptane in the popular Arizona (AZ) liquid system only produced minor changes in the CCC chromatogram with five steroid compounds [19]. Nevertheless, our initial attempts to substitute cyclohexane for n-hexane resulted in unexpected differences in the resulting partitioning coefficients of alkyl hydroxybenzoates. Among the described 20 different solvents, cyclohexane was not listed, suggesting that it was
M. Englert, W. Vetter / J. Chromatogr. A 1342 (2014) 54–62
used in a much smaller extent. Searching for an optimal solvent system for the separation of an alkyl hydroxybenzoate homologous series, cyclohexane was substituted for n-hexane while the amount of other solvents in the system remained unchanged. Following the discovery, it was necessary to further investigate the alkane substitution effect with the employed solvent system. In this work, we therefore used a solvent system based on n-hexane/cyclohexane/tert-butylmethylether/methanol/water in which n-hexane was stepwise substituted by 20% with cyclohexane. In order to evaluate the effects, the different solvent systems were first characterized by physical parameters [13,20], i.e. the exact phase compositions, the densities and polarities of the upper and lower phases, the interfacial tension between the two phases and their viscosity. We also determined exact solvent compositions of the different solvent systems by gas chromatography and Karl-Fischer-Titration. Then, the influence of the n-hexane to cyclohexane ratio was investigated using an alkyl hydroxybenzoate homologous series as easy-to-detect and stable analytes. We determined both, the partitioning coefficients and selectivity factors and then employed the different solvent systems for CCC separations to investigate the selectivity and separation power of the solvent systems. 2. Experimental 2.1. Reagents and materials Ethyl acetate (purest, distilled before use), cyclohexane (>99.5%), methyl-, ethyl-, propyl p-hydroxybenzoate (all 99%) and Hydranal Composite 5 were from Sigma–Aldrich (Steinheim, Germany), methanol (HPLC gradient grade) and tert-butylmethylether (TBME) (purest, distilled before use) were from Fisher Scientific (Leicestshire, Great Britain), nhexane (HPLC gradient grade) was from Th. Geyer (Renningen, Germany), 4 -methoxyacetophenone (99%), 4-(dimethylamino)pyridine (DMAP) (>99%), acetic anhydride (>98%), ethyl m-hydroxybenzoate (99%) and butyl p-hydroxybenzoate (99%) were from Fluka (Taufkirchen, Germany). Isopropyl- and isobutyl p-hydroxybenzoate (both 98%) were from Alfa Aesar (Karlsruhe, Germany). 2.2. Preparation of solvent systems, sample solution and measurement of volume ratios, settling times and physical properties The physico-chemical properties of the solvents used in the examined solvent systems are listed in supplementary data (Table S1). The different two-phase solvent systems used were all compositions of n-hexane and/or cyclohexane as well as methanol, TBME and water. The initial TBME (2) to methanol (5) to water (3) volume ratio was exactly the same for all of the six investigated systems while the combined volume ratio of n-hexane (5-0) and cyclohexane (0-5) resulted in a sum of 5 for the alkane amount. The relative volume proportions of solvents and numbering of the examined n-hexane/cyclohexane/TBME/methanol/water solvent system family are listed in supplementary data (Table S2). The two-phase solvent systems were prepared by adding desired volumes of the solvents into a 2.5 L separatory funnel followed by repeated shaking and equilibration overnight. The content was poured into a glass-stoppered graduated cylinder to measure the volume of the upper and the lower phase. The settling time ts is the time required to form two clear layers after shaking the solvent system [20]. The cylinder containing the two-phase solvent systems was stoppered and settling times were determined by gently turning the cylinder upside down for 5 times.
55
After mixing, the cylinder was immediately placed in the vertical position and the time required for the solvent system to settle into two clear layers was measured. The experiment was repeated 6 times to obtain mean values. Physical properties of the solvent systems including density and viscosity were measured using conventional methods as described below. Densities of the upper and lower phases in each solvent system were determined gravimetrically using a 1 mL pycnometer (Schott Instruments, Mainz, Germany) and an analytical balance UMX2 (Mettler Toledo, Greifensee, Switzerland). Kinematic viscosities of each phase were measured using a routine capillary viscometer (Schott Instruments, Mainz, Germany). All measurements described above were performed at room temperature (22 ± 2 ◦ C) and carried out in duplicate. Before use, the upper and lower phases were separated and degassed for about 10 min in an ultrasonic bath. Sample solutions for the CCC separation were prepared by dissolving 5 mg of the seven alkyl hydroxybenzoates in 2 mL mobile phase of the respective solvent system, resulting in a concentration value of 2.5 mg/mL for each compound or 17.5 mg/mL in total. 2.3. Phase composition determination Exact compositions of n-hexane, cyclohexane, methanol and TBME in the upper organic and the lower aqueous phases of the equilibrated solvent systems were determined by split-injection gas chromatography (GC) coupled with a flame ionization detector (FID) using a Clarus 400 system (Varian, Darmstadt, Germany) in combination with a crossbond diphenyl-/dimethylpolysiloxane Rtx 502.2 column (60 m length, 0.32 mm internal diameter, 1.8 m film thickness; Restek, Bad Homburg, Germany). He (purity 5.0, Westfalengas, Münster, Germany) was used as carrier gas while H2 (purity 5.0, Westfalengas, Münster, Germany) with 45 mL/min and air with 450 mL/min flow rate were used as combustion gases for the FID. The split/splitless injector was operated at 250 ◦ C with a split ratio of 15:1 and injections (0.5 L of upper and lower phase) were performed via the integrated autosampler. Quantification of the solvents was performed by external standard calibration and determination of response factors in multiple standard solutions prepared with a composition similar to that of the solvent samples. Calibration curves were equal to or higher than 0.999 and accurate determination of the solvents was possible. All measurements described above were carried out in triplicate with accuracy estimated to be ± 2%. The water content of both the upper and the lower phases of the equilibrated solvent systems were determined by Karl-FischerTitration with a T70 titrator using a DM143-SC double platinum indicator electrode and a DV705 titration stand (Mettler Toledo, Gießen, Germany). Anhydrous methanol (
Contents lists available at ScienceDirect
Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma
Solvent systems with n-hexane and/or cyclohexane in countercurrent chromatography—Physico-chemical parameters and their impact on the separation of alkyl hydroxybenzoates Michael Englert, Walter Vetter ∗ University of Hohenheim, Institute of Food Chemistry, Garbenstrasse 28, D-70599 Stuttgart, Germany
a r t i c l e
i n f o
Article history: Received 13 January 2014 Received in revised form 28 February 2014 Accepted 19 March 2014 Available online 27 March 2014 Keywords: Countercurrent chromatography Solvents Separation efficiency Alkyl hydroxybenzoates Physico-chemical parameters
a b s t r a c t Countercurrent chromatography (CCC) is an efficient preparative separation technique based on the liquid–liquid distribution of compounds between two phases of a biphasic liquid system. The crucial parameter for the successful application is the selection of the solvent system. Especially for nonpolar analytes the selection options are limited. On the search for a suitable solvent system for the separation of an alkyl hydroxybenzoate homologous series, we noted that the substitution of cyclohexane with n-hexane was accompanied with unexpected differences in partitioning coefficients of the individual analytes. In this study, we investigated the influence of the subsequent substitution of n-hexane with cyclohexane in the n-hexane/cyclohexane/tert-butylmethylether/methanol/water solvent system family. Exact phase compositions and polarity, viscosity and density differences were determined to characterize the different mixtures containing n-hexane and/or cyclohexane. Findings were confirmed by performing CCC separations with different mixtures, which led to baseline resolution for positional isomers when increasing the amount of cyclohexane while the resolution between two pairs of structural isomers decreased. With the new methodology described, structurally similar compounds could be resolved by choosing a certain ratio of n-hexane to cyclohexane. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Countercurrent chromatography (CCC) is a versatile and efficient preparative separation technique for the purification of various natural [1–5] and synthetic products [6–8] from crude mixtures as well as for the isolation of pharmaceutical drugs [9,10]. The chromatographic separation is based on the liquid–liquid distribution of analytes between mobile and stationary phases which represent two immiscible liquids [11,12]. Centrifugal fields are used to maintain the stationary phase in the column while the mobile phase is pumped through it to elute the analytes according to their partitioning coefficients (K) [13]. The selection of an appropriate solvent system is of utmost importance for the successful application and comparable to the role of both, the column and an eluent in high-performance liquid chromatography (HPLC) [12,14,15]. Numerous examples of twophase solvent systems are described that have been successfully employed for separations over the years in the literature [14,15].
∗ Corresponding author. Tel.: +49 711 459 24016; fax: +49 711 459 24377. E-mail address: [email protected] (W. Vetter). http://dx.doi.org/10.1016/j.chroma.2014.03.050 0021-9673/© 2014 Elsevier B.V. All rights reserved.
In general, three important criteria should be fulfilled by the solvent system for successful CCC separations. First, it should form two clear distinct phases with a separation time (ts ) less than 30 s to preclude emulsification [11,16]. Second, the selected solvent system should provide a high retention of the stationary phase, measured by the stationary phase retention (Sf ) [11]. Third, the sample compounds should have partitioning coefficients K between 0.4 and 2.5 with significant differences between the individual analytes, corresponding to high chromatographic selectivity [13,17]. As reported in a recent review, only a restricted number of solvents are employed in CCC with water, n-hexane, ethyl acetate and methanol as the top 4 utilitarian solvents [18]. Especially for nonpolar analytes the choice of solvent mixtures is very limited. The less polar solvent is usually represented by an alkane. Berthod et al. demonstrated that different alkanes (n-pentane, n-hexane, cyclohexane and iso-octane) used in place of n-heptane in the popular Arizona (AZ) liquid system only produced minor changes in the CCC chromatogram with five steroid compounds [19]. Nevertheless, our initial attempts to substitute cyclohexane for n-hexane resulted in unexpected differences in the resulting partitioning coefficients of alkyl hydroxybenzoates. Among the described 20 different solvents, cyclohexane was not listed, suggesting that it was
M. Englert, W. Vetter / J. Chromatogr. A 1342 (2014) 54–62
used in a much smaller extent. Searching for an optimal solvent system for the separation of an alkyl hydroxybenzoate homologous series, cyclohexane was substituted for n-hexane while the amount of other solvents in the system remained unchanged. Following the discovery, it was necessary to further investigate the alkane substitution effect with the employed solvent system. In this work, we therefore used a solvent system based on n-hexane/cyclohexane/tert-butylmethylether/methanol/water in which n-hexane was stepwise substituted by 20% with cyclohexane. In order to evaluate the effects, the different solvent systems were first characterized by physical parameters [13,20], i.e. the exact phase compositions, the densities and polarities of the upper and lower phases, the interfacial tension between the two phases and their viscosity. We also determined exact solvent compositions of the different solvent systems by gas chromatography and Karl-Fischer-Titration. Then, the influence of the n-hexane to cyclohexane ratio was investigated using an alkyl hydroxybenzoate homologous series as easy-to-detect and stable analytes. We determined both, the partitioning coefficients and selectivity factors and then employed the different solvent systems for CCC separations to investigate the selectivity and separation power of the solvent systems. 2. Experimental 2.1. Reagents and materials Ethyl acetate (purest, distilled before use), cyclohexane (>99.5%), methyl-, ethyl-, propyl p-hydroxybenzoate (all 99%) and Hydranal Composite 5 were from Sigma–Aldrich (Steinheim, Germany), methanol (HPLC gradient grade) and tert-butylmethylether (TBME) (purest, distilled before use) were from Fisher Scientific (Leicestshire, Great Britain), nhexane (HPLC gradient grade) was from Th. Geyer (Renningen, Germany), 4 -methoxyacetophenone (99%), 4-(dimethylamino)pyridine (DMAP) (>99%), acetic anhydride (>98%), ethyl m-hydroxybenzoate (99%) and butyl p-hydroxybenzoate (99%) were from Fluka (Taufkirchen, Germany). Isopropyl- and isobutyl p-hydroxybenzoate (both 98%) were from Alfa Aesar (Karlsruhe, Germany). 2.2. Preparation of solvent systems, sample solution and measurement of volume ratios, settling times and physical properties The physico-chemical properties of the solvents used in the examined solvent systems are listed in supplementary data (Table S1). The different two-phase solvent systems used were all compositions of n-hexane and/or cyclohexane as well as methanol, TBME and water. The initial TBME (2) to methanol (5) to water (3) volume ratio was exactly the same for all of the six investigated systems while the combined volume ratio of n-hexane (5-0) and cyclohexane (0-5) resulted in a sum of 5 for the alkane amount. The relative volume proportions of solvents and numbering of the examined n-hexane/cyclohexane/TBME/methanol/water solvent system family are listed in supplementary data (Table S2). The two-phase solvent systems were prepared by adding desired volumes of the solvents into a 2.5 L separatory funnel followed by repeated shaking and equilibration overnight. The content was poured into a glass-stoppered graduated cylinder to measure the volume of the upper and the lower phase. The settling time ts is the time required to form two clear layers after shaking the solvent system [20]. The cylinder containing the two-phase solvent systems was stoppered and settling times were determined by gently turning the cylinder upside down for 5 times.
55
After mixing, the cylinder was immediately placed in the vertical position and the time required for the solvent system to settle into two clear layers was measured. The experiment was repeated 6 times to obtain mean values. Physical properties of the solvent systems including density and viscosity were measured using conventional methods as described below. Densities of the upper and lower phases in each solvent system were determined gravimetrically using a 1 mL pycnometer (Schott Instruments, Mainz, Germany) and an analytical balance UMX2 (Mettler Toledo, Greifensee, Switzerland). Kinematic viscosities of each phase were measured using a routine capillary viscometer (Schott Instruments, Mainz, Germany). All measurements described above were performed at room temperature (22 ± 2 ◦ C) and carried out in duplicate. Before use, the upper and lower phases were separated and degassed for about 10 min in an ultrasonic bath. Sample solutions for the CCC separation were prepared by dissolving 5 mg of the seven alkyl hydroxybenzoates in 2 mL mobile phase of the respective solvent system, resulting in a concentration value of 2.5 mg/mL for each compound or 17.5 mg/mL in total. 2.3. Phase composition determination Exact compositions of n-hexane, cyclohexane, methanol and TBME in the upper organic and the lower aqueous phases of the equilibrated solvent systems were determined by split-injection gas chromatography (GC) coupled with a flame ionization detector (FID) using a Clarus 400 system (Varian, Darmstadt, Germany) in combination with a crossbond diphenyl-/dimethylpolysiloxane Rtx 502.2 column (60 m length, 0.32 mm internal diameter, 1.8 m film thickness; Restek, Bad Homburg, Germany). He (purity 5.0, Westfalengas, Münster, Germany) was used as carrier gas while H2 (purity 5.0, Westfalengas, Münster, Germany) with 45 mL/min and air with 450 mL/min flow rate were used as combustion gases for the FID. The split/splitless injector was operated at 250 ◦ C with a split ratio of 15:1 and injections (0.5 L of upper and lower phase) were performed via the integrated autosampler. Quantification of the solvents was performed by external standard calibration and determination of response factors in multiple standard solutions prepared with a composition similar to that of the solvent samples. Calibration curves were equal to or higher than 0.999 and accurate determination of the solvents was possible. All measurements described above were carried out in triplicate with accuracy estimated to be ± 2%. The water content of both the upper and the lower phases of the equilibrated solvent systems were determined by Karl-FischerTitration with a T70 titrator using a DM143-SC double platinum indicator electrode and a DV705 titration stand (Mettler Toledo, Gießen, Germany). Anhydrous methanol (
