- Email: [email protected]
Capillary liquid chromatography and capillary electrochromatography using silica hydride stationary phases
Available online at www.sciencedirect.com
Journal of Chromatography A, 1191 (2008) 136–140
Capillary liquid chromatography and capillary electrochromatography using silica hydride stationary phases Joseph J. Pesek ∗ , Maria T. Matyska, Dipti Sukul Department of Chemistry, San Jose State University, One Washington Square, San Jose, CA 95112, USA Available online 8 February 2008
Abstract A hydride-based octadecyl stationary phase on both 4.0 and 1.8 m silica particles is tested in both the capillary LC and the pressurized capillary electrochromatography (pCEC) modes. These two materials are compared to standard C18 stationary phase made by organosilanization and to the hydride material packed into a convention 4.6 mm I.D. column. The performance of the capillary columns is evaluated in terms of analysis times for various mixtures as well as efficiency. Of particular interest are the differences between the LC mode where only laminar flow is present and pCEC operation where a flat electrodriven flow profile is superimposed on the parabolic pressurized flow. Differences in performance between columns packed with 4.0 and 1.8 m particle silica are also evaluated. © 2008 Elsevier B.V. All rights reserved. Keywords: Hydrosilation; Sub-2 m silica particles; Pressurized CEC
1. Introduction In general, all chemically bonded stationary phases based on silica are fabricated using one of two organosilanization reaction processes [1,2]. The first reaction utilizes a trifunctional organosilane to react with the silica surface and produces what is referred to as a polymeric stationary phase [1–3]. In this case the bonded material crosslinks leading to a polymeric network on the surface. In many cases this reaction is difficult to control and hence the amount of bonded material on the surface can be variable. The second reaction uses a monofunctional organosilane with a single point of attachment to the silica surface. The monofunctional organosilane reactions are generally more reproducible than the polymeric process so that this approach is used more often for commercial materials. In the second reaction, two of the positions on the organosilane have methyl groups but other moieties have been used for other purposes such as steric blocking of the surface [4]. For both reactions the underlying silica still retains a rather large number of silanols. In the case of a monomeric phase, steric considerations involving the surface and size of the monofunctional organosilane preclude having more than about 50% of the original silanols covered; in some bonded phases for large organic groups on the silane this ∗
Corresponding author. Fax: +1 4089244945. E-mail address: [email protected] (J.J. Pesek).
0021-9673/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2008.02.006
value is much lower. Polymeric phases have the same limitation plus the possibility of incomplete crosslinking that would lead to additional silanols being produced. On some phases the residual silanols are allowed to remain but for some materials the number is reduced by secondary reactions referred to as endcapping [2,5–7]. Another method has been developed over the last decade that creates a new material based on silica that has very few silanols on the surface. This new support material is silica hydride where Si–H moieties have replaced approximately 95% of the Si–OH groups on the silica surface [8,9]. One reaction protocol for creating a hydride surface and then modifying it with different organic groups is silanization/hydrosilation [10–12]. The first step produces a hydride surface and then the second step attaches the desired organic functionality for the stationary phase. If an alkene is used in the hydrosilation a monodentate bonded phase is formed.
Double attachment of the bonded group to the surface occurs when an alkyne is used in the hydrosilation reaction [13,14]. For both the monodentate and bidentate materials, the underlying structure of the surface is different because the unreacted sites on the surface are Si–H groups rather than the silanols that are present on stationary phases fabricated by organosilanization on
J.J. Pesek et al. / J. Chromatogr. A 1191 (2008) 136–140
137
ordinary silica. The presence of silicon hydrides as opposed to the silanols leads to different and in many cases desirable properties for HPLC stationary phases [8,9,15]. The same approach to surface modification has been developed for fused capillaries used in capillary electrochromatography (CEC). In this investigation, particulate silica (both 4.2 and 1.8 m) is modified by the silanization/hydrosilation procedures. The modified materials are then packed into fused silica capillary tubes and tested in both the -HPLC and CEC modes. Such a use of hydride-based stationary phases has not been studied to date. 2. Experimental 2.1. Materials The hydride-based stationary phase columns used in this study were obtained from Microsolv Technology (Eatontown, NJ, USA). The silica hydride stationary phase was a bidentate C18 material on either 4.0 or 1.8 m silica particles. For 4.0 m particles the capillary dimensions were 100 m I.D. with a 20 cm packed length and a total length of 30 cm. The 1.8 m particle column had an I.D. of 50 m with a packed length of 20 cm and a total length of 35 cm. The conventional C18 bonded phase column was obtained from Unimicro Technologies (Pleasanton, CA, USA). The particle diameter was 5 m, the column I.D. was 100 m and the total length was 30 cm with the packed length being 20 cm. All columns were packed by Unimicro Technologies using an electrokinetic method [16]. The solutes used were purchased from Sigma–Aldrich (Milwaukee, WI, USA). The organic components for the HPLC mobile phases were all obtained from Fisher Scientific, Fair Lawn, NJ, USA. Deionized water was prepared on a Milli-Q water purification system (Millipore, Bedford, MA, USA). 2.2. Instrumentation The integrated capillary liquid chromatography and capillary electrochromatography instrument was a Unimicro Technologies TriSep Model 2100 and was equipped with dual micro LC pumps, a manual injection system, a high voltage power supply for pCEC operation and a variable wavelength UV detector. The standard injection volume for all samples as determined by the six-port nanovalve of the TriSep system was 10 nL. 3. Results and discussion Hydride-based stationary phases have unique properties in comparison to most materials fabricated by organosilanization. By manipulation of the mobile phase composition, a stationary phase with a hydride surface can function in any of the following three separation modes: reversed-phase; aqueous normal-phase and organic normal-phase [14]. In this study both the capillary liquid chromatography and capillary electrochromatographic properties of these materials are investigated demonstrating reversed phase and aqueous normal phase retention.
Fig. 1. Separation of steroid mixture on C18 columns. (A) Capillary column with 5.0 m stationary phase made by organosilanization. (B) Capillary column with 4.0 m hydride-based stationary phase. Mobile phase 80:20 acetonitrile/DI water (0.1% formic acid). Detection at 254 nm. Flow rate = 70 L/min. Pressure 1022 psi. Solutes: (1) Prednisolone; (2) corticosterone; (3) norgestrel; (4) progesterone.
In the reversed-phase mode, a stationary phase having a C18 moiety bonded to hydride surface should behave similarly to ordinary silica materials modified by organosilanization when the solutes are typical hydrophobic compounds. Fig. 1A shows the chromatogram for a mixture of four steroids using a standard C18 phase in a -HPLC system. This result can be compared to the separation obtained on a hydride C18 column under the same experimental conditions (Fig. 1B). While both bonded materials have an octadecyl moiety bonded to the surface, there are some noticeable differences in the two chromatograms. The elution order is the same with the most polar compound (prednisolone) having three hydroxyl groups and two carbonyls as the first peak followed by corticosterone (two hydroxyls and two carbonyls), norgestrel (one hydroxyl and two carbonyls) and finally progesterone (two carbonyls). The columns have the same dimensions and the experimental conditions were identical. The first two peaks are retained longer on the column produced
138
J.J. Pesek et al. / J. Chromatogr. A 1191 (2008) 136–140
by organosilanization while the third peak has approximately the same retention on the two columns and the last analyte is retained longer on the hydride-based column. The other noticeable difference is the tailing observed for component four on the organsilanization column while this compound has significantly better symmetry on the hydride column. In order to verify the peak shape of progesterone on the bidentate C18 (hydride) column, the same mixture was run under standard HPLC conditions (4.6 mm I.D. × 75 mm column) with this stationary phase. The peak shape for progesterone has the same good symmetry as in the capillary configuration. As expected, plots of retention time vs. % acetonitrile (ACN) in the mobile phase show decreasing retention with increasing amounts of ACN (reversed-phase behavior). In order to determine the effect of voltage (pCEC) on the migration of the steroids, a series of experiments at 5, 10, 15 and 20 kV using several different acetonitrile compositions were run. A typical result is shown in Fig. 2 for the separation of the four component steroid mixture at 80% ACN and an applied field of 20 kV. In comparison to Fig. 1B, it can be seen that the elution time for all of the components has been shortened considerably (4.5 min vs. 11.0 min for the last peak). This is due to the presence of electroosmotic flow (EOF) that provides faster movement of the mobile phase through the column. The EOF is generated by the silanols below the particulate silica hydride surface as well as the SiOH groups on the fused silica capillary wall. The overall movement of solutes through the column then is a combination of laminar flow due to the applied pressure and the EOF generated by the applied field. The latter has a flat flow
Fig. 2. Separation of steroid mixture by pCEC on column packed with 4.0 m hydride-based C18. Applied voltage: 20 kV. Other conditions and solutes the same as Fig. 1.
profile while the pressure driven flow is laminar. The pCEC then results in a hybrid flow profile that should lead to some improvement in efficiency. Indeed, a small improvement is observed for norgestrel (65,000 plates/m at 20 kV vs. 58,000 plates/m at 0 kV) and progesterone (68,000 plates/m at 20 kV vs. 60,000 plates/m at 0 kV). A plot of the migration time vs. the applied voltage for the four steroid compounds shows a general trend of shorter elution time with increasing applied voltage. A dramatic improvement in the speed of analysis and efficiency can be obtained for charged compounds on the hydride-based C18 column when comparing the results for HPLC and pCEC. In Fig. 3A a mixture of guanosine, thymine and cytidine is run under pressure driven elution. The total analysis time is approximately 12 min and the third peak is somewhat
Fig. 3. Separation of a mixture containing two nucleobases and a nucleoside. (A) by capillary LC and (B) by pCEC. Column: 4.0 m hydride-based bidentate C18. Mobile phase 60:40 acetonitrile/DI water (0.1% formic acid). Flow rate = 70 L/min. Applied voltage in B = 10 kV. Detection at 254 nm. Solutes: (1) guanosine; (2) thymine; (3) cytidine.
J.J. Pesek et al. / J. Chromatogr. A 1191 (2008) 136–140
Fig. 4. Retention map (retention time vs. % acetonitrile in mobile phase) for three drugs using the hydride-based bidentate C18 column.
asymmetric. When only 10 kV is applied to the system (Fig. 3B), the analysis time is reduced to about 5 min and the shape of the third peak is dramatically improved. In this case, the molecules in the mixture are driven through the column by three mechanisms: the laminar flow provided by the pumps; the EOF created in the capillary by the electrolyte; and the electrophoretic mobility of the charged analytes. The effect of electrophoretic mobility is evident by the fact that when an electric field is applied, two of the solutes change position in the elution order. At the pH of the mobile phase (∼3.6) cytidine has the highest effective positive charge and hence its elution time is decreased more than the other two solutes. One of the unique properties of the hydride-based stationary phases is their ability to operate in both the reversed-phase and aqueous normal-phase modes [14]. Thus samples that are both hydrophilic and hydrophobic can be retained by the same stationary phase, often under the same mobile phase conditions. To evaluate this behavior, the retention of three drugs (amphotericin, metformin and sulfonamide) was tested as a function of mobile phase composition. Fig. 4 is the retention map for these three compounds on the bidentate C18 hydride phase. In this case, two of the drugs exhibited aqueous normal-phase behavior, i.e. increasing retention as the amount of the least polar component (acetonitrile) in the mobile phase is increased. These results are consistent with those obtained on a standard analytical HPLC column packed with a hydride-based C18 stationary phase [17] and are consistent with the relative polarity of the molecules. Analogous to the properties observed in the reversed-phase, when a voltage is applied to samples containing hydrophilic compounds the analysis time decreases in the aqueous normal phase. Recently there has been considerable interest in the use of sub-2 m particles for stationary phases in HPLC. The main advantages are in higher efficiencies and faster analysis times when using columns with stationary phases having smaller particle diameters. Since the pressure capabilities of the system in use are not sufficient for true ultra-performance liquid chromatography (UPLC) operation, this aspect of small particles in the capillary LC mode could not be fully tested. Under the same pressure and mobile phase conditions shown in Fig. 1, the efficiencies for norgestrel and progesterone were approximately 80,000 plates/m when using the 1.8 m particle column.
139
Fig. 5. Electrochromatogram for pCEC of hormone mixture on the 1.8 m hydride-based bidentate C18 column. Mobile phase 60:40 acetonitrile/DI water (0.1% formic acid). Applied voltage = 20 kV. Other conditions same as Fig. 1.
The retention maps for each of the analytes are similar to those obtained for the 4.0 m particles with typical reversed-phase behavior. A comparison of the relative efficiencies in the pCEC under comparable conditions can also be made in order to determine the impact of the smaller particles under combined laminar and electrodriven flow conditions. The chromatogram for the four steroid mixture is shown in Fig. 5. In this case, a more dramatic improvement in column efficiency is obtained with all components having in excess of 200,000 plates/m. As described, the steroid migration time is reduced significantly in pCEC operation in comparison to when no voltage is applied to the system. A plot of migration time as a function of applied voltage for a number of charged solutes shows that retention is reduced to a greater extent due to the presence of electrophoretic mobility as well as EOF. Thus even without the application of the extremely high pressures typically found in UPLC operation, the use of moderate pressure in combination with electroosmotic flow as well as electrophoretic migration for charged analytes will yield both high efficiencies and significantly reduced analysis times. 4. Conclusions Hydride-based stationary phases are useful separation media for both capillary LC and pCEC operation. It has been demonstrated that the hydride-based C18 can operate in both the reversed-phase and aqueous normal-phase modes. The parameter connecting these two modes is the presence of water in the mobile with high aqueous percentages resulting in reversedphase behavior for hydrophobic compounds and high organic concentrations leading to aqueous normal-phase retention for hydrophilic compounds. For one mixture of compounds containing both polar and nonpolar species, the two modes overlap at intermediate mobile phase compositions. The use of an applied voltage to the system in addition to pressure speeds up the analysis and improves efficiencies for both neutral and charged compounds. Acknowledgements The financial support of the National Institutes of Health (Grant GM R15 GM079741-01) and the loan of the capillary
140
J.J. Pesek et al. / J. Chromatogr. A 1191 (2008) 136–140
LC/CEC instrument by Unimicro Technologies are gratefully acknowledged. J.J.P. would like to acknowledge the support of the Camille and Henry Dreyfus Foundation through a Scholar Award.
[6] [7] [8] [9] [10]
References
[11]
[1] E.F. Vansant, P. Van Der Voort, K.C. Vrancken, Characterization and Chemical Modification of Silica, Elsevier, Amsterdam, 1995. [2] J. Nawrocki, J. Chromatogr. A 779 (1997) 29. [3] L.C. Sander, S.A. Wise, CRC Crit. Rev. Anal. Chem. 18 (1987) 299. [4] J.J. Kirkland, M.A. van Straten, H.A. Claessens, J. Chromatogr. A 691 (1995) 3. [5] C.H. Lochmuller, D.B. Marshall, Anal. Chim. Acta 142 (1982) 63.
[12] [13] [14] [15] [16] [17]
D.B. Marshall, C.H. Lochmuller, J. Chromatogr. Sci. 25 (1987) 262. L.C. Sander, S.A. Wise, Crit. Rev. Anal. Chem. 18 (1987) 299. J.J. Pesek, M.T. Matyska, J. Sep. Sci. 28 (2005) 1845. J.J. Pesek, M.T. Matyska, J. Liq. Chromatogr. Rel. Technol. 29 (2006) 1105. J.J. Pesek, M.T. Matyska, J.E. Sandoval, E. Williamsen, J. Liq. Chromatogr. Rel. Technol. 19 (1996) 2843. C.H. Chu, E. Jonsson, M. Auvinen, J.J. Pesek, J.E. Sandoval, Anal. Chem. 65 (1993) 808. J.E. Sandoval, J.J. Pesek, Anal. Chem. 63 (1991) 2634. J.J. Pesek, M.T. Matyska, M. Oliva, M. Evanchic, J. Chromatogr. A 818 (1998) 145. J.J. Pesek, M.T. Matyska, S. Prabhakaran, J. Sep. Sci. 28 (2005) 2437. J.J. Pesek, M.T. Matyska, LC/GC 24 (2006) 296. C. Yan, US Patent 5,453,763 (1995). J.J. Pesek, M.T. Matyska, unpublished results.
Journal of Chromatography A, 1191 (2008) 136–140
Capillary liquid chromatography and capillary electrochromatography using silica hydride stationary phases Joseph J. Pesek ∗ , Maria T. Matyska, Dipti Sukul Department of Chemistry, San Jose State University, One Washington Square, San Jose, CA 95112, USA Available online 8 February 2008
Abstract A hydride-based octadecyl stationary phase on both 4.0 and 1.8 m silica particles is tested in both the capillary LC and the pressurized capillary electrochromatography (pCEC) modes. These two materials are compared to standard C18 stationary phase made by organosilanization and to the hydride material packed into a convention 4.6 mm I.D. column. The performance of the capillary columns is evaluated in terms of analysis times for various mixtures as well as efficiency. Of particular interest are the differences between the LC mode where only laminar flow is present and pCEC operation where a flat electrodriven flow profile is superimposed on the parabolic pressurized flow. Differences in performance between columns packed with 4.0 and 1.8 m particle silica are also evaluated. © 2008 Elsevier B.V. All rights reserved. Keywords: Hydrosilation; Sub-2 m silica particles; Pressurized CEC
1. Introduction In general, all chemically bonded stationary phases based on silica are fabricated using one of two organosilanization reaction processes [1,2]. The first reaction utilizes a trifunctional organosilane to react with the silica surface and produces what is referred to as a polymeric stationary phase [1–3]. In this case the bonded material crosslinks leading to a polymeric network on the surface. In many cases this reaction is difficult to control and hence the amount of bonded material on the surface can be variable. The second reaction uses a monofunctional organosilane with a single point of attachment to the silica surface. The monofunctional organosilane reactions are generally more reproducible than the polymeric process so that this approach is used more often for commercial materials. In the second reaction, two of the positions on the organosilane have methyl groups but other moieties have been used for other purposes such as steric blocking of the surface [4]. For both reactions the underlying silica still retains a rather large number of silanols. In the case of a monomeric phase, steric considerations involving the surface and size of the monofunctional organosilane preclude having more than about 50% of the original silanols covered; in some bonded phases for large organic groups on the silane this ∗
Corresponding author. Fax: +1 4089244945. E-mail address: [email protected] (J.J. Pesek).
0021-9673/$ – see front matter © 2008 Elsevier B.V. All rights reserved. doi:10.1016/j.chroma.2008.02.006
value is much lower. Polymeric phases have the same limitation plus the possibility of incomplete crosslinking that would lead to additional silanols being produced. On some phases the residual silanols are allowed to remain but for some materials the number is reduced by secondary reactions referred to as endcapping [2,5–7]. Another method has been developed over the last decade that creates a new material based on silica that has very few silanols on the surface. This new support material is silica hydride where Si–H moieties have replaced approximately 95% of the Si–OH groups on the silica surface [8,9]. One reaction protocol for creating a hydride surface and then modifying it with different organic groups is silanization/hydrosilation [10–12]. The first step produces a hydride surface and then the second step attaches the desired organic functionality for the stationary phase. If an alkene is used in the hydrosilation a monodentate bonded phase is formed.
Double attachment of the bonded group to the surface occurs when an alkyne is used in the hydrosilation reaction [13,14]. For both the monodentate and bidentate materials, the underlying structure of the surface is different because the unreacted sites on the surface are Si–H groups rather than the silanols that are present on stationary phases fabricated by organosilanization on
J.J. Pesek et al. / J. Chromatogr. A 1191 (2008) 136–140
137
ordinary silica. The presence of silicon hydrides as opposed to the silanols leads to different and in many cases desirable properties for HPLC stationary phases [8,9,15]. The same approach to surface modification has been developed for fused capillaries used in capillary electrochromatography (CEC). In this investigation, particulate silica (both 4.2 and 1.8 m) is modified by the silanization/hydrosilation procedures. The modified materials are then packed into fused silica capillary tubes and tested in both the -HPLC and CEC modes. Such a use of hydride-based stationary phases has not been studied to date. 2. Experimental 2.1. Materials The hydride-based stationary phase columns used in this study were obtained from Microsolv Technology (Eatontown, NJ, USA). The silica hydride stationary phase was a bidentate C18 material on either 4.0 or 1.8 m silica particles. For 4.0 m particles the capillary dimensions were 100 m I.D. with a 20 cm packed length and a total length of 30 cm. The 1.8 m particle column had an I.D. of 50 m with a packed length of 20 cm and a total length of 35 cm. The conventional C18 bonded phase column was obtained from Unimicro Technologies (Pleasanton, CA, USA). The particle diameter was 5 m, the column I.D. was 100 m and the total length was 30 cm with the packed length being 20 cm. All columns were packed by Unimicro Technologies using an electrokinetic method [16]. The solutes used were purchased from Sigma–Aldrich (Milwaukee, WI, USA). The organic components for the HPLC mobile phases were all obtained from Fisher Scientific, Fair Lawn, NJ, USA. Deionized water was prepared on a Milli-Q water purification system (Millipore, Bedford, MA, USA). 2.2. Instrumentation The integrated capillary liquid chromatography and capillary electrochromatography instrument was a Unimicro Technologies TriSep Model 2100 and was equipped with dual micro LC pumps, a manual injection system, a high voltage power supply for pCEC operation and a variable wavelength UV detector. The standard injection volume for all samples as determined by the six-port nanovalve of the TriSep system was 10 nL. 3. Results and discussion Hydride-based stationary phases have unique properties in comparison to most materials fabricated by organosilanization. By manipulation of the mobile phase composition, a stationary phase with a hydride surface can function in any of the following three separation modes: reversed-phase; aqueous normal-phase and organic normal-phase [14]. In this study both the capillary liquid chromatography and capillary electrochromatographic properties of these materials are investigated demonstrating reversed phase and aqueous normal phase retention.
Fig. 1. Separation of steroid mixture on C18 columns. (A) Capillary column with 5.0 m stationary phase made by organosilanization. (B) Capillary column with 4.0 m hydride-based stationary phase. Mobile phase 80:20 acetonitrile/DI water (0.1% formic acid). Detection at 254 nm. Flow rate = 70 L/min. Pressure 1022 psi. Solutes: (1) Prednisolone; (2) corticosterone; (3) norgestrel; (4) progesterone.
In the reversed-phase mode, a stationary phase having a C18 moiety bonded to hydride surface should behave similarly to ordinary silica materials modified by organosilanization when the solutes are typical hydrophobic compounds. Fig. 1A shows the chromatogram for a mixture of four steroids using a standard C18 phase in a -HPLC system. This result can be compared to the separation obtained on a hydride C18 column under the same experimental conditions (Fig. 1B). While both bonded materials have an octadecyl moiety bonded to the surface, there are some noticeable differences in the two chromatograms. The elution order is the same with the most polar compound (prednisolone) having three hydroxyl groups and two carbonyls as the first peak followed by corticosterone (two hydroxyls and two carbonyls), norgestrel (one hydroxyl and two carbonyls) and finally progesterone (two carbonyls). The columns have the same dimensions and the experimental conditions were identical. The first two peaks are retained longer on the column produced
138
J.J. Pesek et al. / J. Chromatogr. A 1191 (2008) 136–140
by organosilanization while the third peak has approximately the same retention on the two columns and the last analyte is retained longer on the hydride-based column. The other noticeable difference is the tailing observed for component four on the organsilanization column while this compound has significantly better symmetry on the hydride column. In order to verify the peak shape of progesterone on the bidentate C18 (hydride) column, the same mixture was run under standard HPLC conditions (4.6 mm I.D. × 75 mm column) with this stationary phase. The peak shape for progesterone has the same good symmetry as in the capillary configuration. As expected, plots of retention time vs. % acetonitrile (ACN) in the mobile phase show decreasing retention with increasing amounts of ACN (reversed-phase behavior). In order to determine the effect of voltage (pCEC) on the migration of the steroids, a series of experiments at 5, 10, 15 and 20 kV using several different acetonitrile compositions were run. A typical result is shown in Fig. 2 for the separation of the four component steroid mixture at 80% ACN and an applied field of 20 kV. In comparison to Fig. 1B, it can be seen that the elution time for all of the components has been shortened considerably (4.5 min vs. 11.0 min for the last peak). This is due to the presence of electroosmotic flow (EOF) that provides faster movement of the mobile phase through the column. The EOF is generated by the silanols below the particulate silica hydride surface as well as the SiOH groups on the fused silica capillary wall. The overall movement of solutes through the column then is a combination of laminar flow due to the applied pressure and the EOF generated by the applied field. The latter has a flat flow
Fig. 2. Separation of steroid mixture by pCEC on column packed with 4.0 m hydride-based C18. Applied voltage: 20 kV. Other conditions and solutes the same as Fig. 1.
profile while the pressure driven flow is laminar. The pCEC then results in a hybrid flow profile that should lead to some improvement in efficiency. Indeed, a small improvement is observed for norgestrel (65,000 plates/m at 20 kV vs. 58,000 plates/m at 0 kV) and progesterone (68,000 plates/m at 20 kV vs. 60,000 plates/m at 0 kV). A plot of the migration time vs. the applied voltage for the four steroid compounds shows a general trend of shorter elution time with increasing applied voltage. A dramatic improvement in the speed of analysis and efficiency can be obtained for charged compounds on the hydride-based C18 column when comparing the results for HPLC and pCEC. In Fig. 3A a mixture of guanosine, thymine and cytidine is run under pressure driven elution. The total analysis time is approximately 12 min and the third peak is somewhat
Fig. 3. Separation of a mixture containing two nucleobases and a nucleoside. (A) by capillary LC and (B) by pCEC. Column: 4.0 m hydride-based bidentate C18. Mobile phase 60:40 acetonitrile/DI water (0.1% formic acid). Flow rate = 70 L/min. Applied voltage in B = 10 kV. Detection at 254 nm. Solutes: (1) guanosine; (2) thymine; (3) cytidine.
J.J. Pesek et al. / J. Chromatogr. A 1191 (2008) 136–140
Fig. 4. Retention map (retention time vs. % acetonitrile in mobile phase) for three drugs using the hydride-based bidentate C18 column.
asymmetric. When only 10 kV is applied to the system (Fig. 3B), the analysis time is reduced to about 5 min and the shape of the third peak is dramatically improved. In this case, the molecules in the mixture are driven through the column by three mechanisms: the laminar flow provided by the pumps; the EOF created in the capillary by the electrolyte; and the electrophoretic mobility of the charged analytes. The effect of electrophoretic mobility is evident by the fact that when an electric field is applied, two of the solutes change position in the elution order. At the pH of the mobile phase (∼3.6) cytidine has the highest effective positive charge and hence its elution time is decreased more than the other two solutes. One of the unique properties of the hydride-based stationary phases is their ability to operate in both the reversed-phase and aqueous normal-phase modes [14]. Thus samples that are both hydrophilic and hydrophobic can be retained by the same stationary phase, often under the same mobile phase conditions. To evaluate this behavior, the retention of three drugs (amphotericin, metformin and sulfonamide) was tested as a function of mobile phase composition. Fig. 4 is the retention map for these three compounds on the bidentate C18 hydride phase. In this case, two of the drugs exhibited aqueous normal-phase behavior, i.e. increasing retention as the amount of the least polar component (acetonitrile) in the mobile phase is increased. These results are consistent with those obtained on a standard analytical HPLC column packed with a hydride-based C18 stationary phase [17] and are consistent with the relative polarity of the molecules. Analogous to the properties observed in the reversed-phase, when a voltage is applied to samples containing hydrophilic compounds the analysis time decreases in the aqueous normal phase. Recently there has been considerable interest in the use of sub-2 m particles for stationary phases in HPLC. The main advantages are in higher efficiencies and faster analysis times when using columns with stationary phases having smaller particle diameters. Since the pressure capabilities of the system in use are not sufficient for true ultra-performance liquid chromatography (UPLC) operation, this aspect of small particles in the capillary LC mode could not be fully tested. Under the same pressure and mobile phase conditions shown in Fig. 1, the efficiencies for norgestrel and progesterone were approximately 80,000 plates/m when using the 1.8 m particle column.
139
Fig. 5. Electrochromatogram for pCEC of hormone mixture on the 1.8 m hydride-based bidentate C18 column. Mobile phase 60:40 acetonitrile/DI water (0.1% formic acid). Applied voltage = 20 kV. Other conditions same as Fig. 1.
The retention maps for each of the analytes are similar to those obtained for the 4.0 m particles with typical reversed-phase behavior. A comparison of the relative efficiencies in the pCEC under comparable conditions can also be made in order to determine the impact of the smaller particles under combined laminar and electrodriven flow conditions. The chromatogram for the four steroid mixture is shown in Fig. 5. In this case, a more dramatic improvement in column efficiency is obtained with all components having in excess of 200,000 plates/m. As described, the steroid migration time is reduced significantly in pCEC operation in comparison to when no voltage is applied to the system. A plot of migration time as a function of applied voltage for a number of charged solutes shows that retention is reduced to a greater extent due to the presence of electrophoretic mobility as well as EOF. Thus even without the application of the extremely high pressures typically found in UPLC operation, the use of moderate pressure in combination with electroosmotic flow as well as electrophoretic migration for charged analytes will yield both high efficiencies and significantly reduced analysis times. 4. Conclusions Hydride-based stationary phases are useful separation media for both capillary LC and pCEC operation. It has been demonstrated that the hydride-based C18 can operate in both the reversed-phase and aqueous normal-phase modes. The parameter connecting these two modes is the presence of water in the mobile with high aqueous percentages resulting in reversedphase behavior for hydrophobic compounds and high organic concentrations leading to aqueous normal-phase retention for hydrophilic compounds. For one mixture of compounds containing both polar and nonpolar species, the two modes overlap at intermediate mobile phase compositions. The use of an applied voltage to the system in addition to pressure speeds up the analysis and improves efficiencies for both neutral and charged compounds. Acknowledgements The financial support of the National Institutes of Health (Grant GM R15 GM079741-01) and the loan of the capillary
140
J.J. Pesek et al. / J. Chromatogr. A 1191 (2008) 136–140
LC/CEC instrument by Unimicro Technologies are gratefully acknowledged. J.J.P. would like to acknowledge the support of the Camille and Henry Dreyfus Foundation through a Scholar Award.
[6] [7] [8] [9] [10]
References
[11]
[1] E.F. Vansant, P. Van Der Voort, K.C. Vrancken, Characterization and Chemical Modification of Silica, Elsevier, Amsterdam, 1995. [2] J. Nawrocki, J. Chromatogr. A 779 (1997) 29. [3] L.C. Sander, S.A. Wise, CRC Crit. Rev. Anal. Chem. 18 (1987) 299. [4] J.J. Kirkland, M.A. van Straten, H.A. Claessens, J. Chromatogr. A 691 (1995) 3. [5] C.H. Lochmuller, D.B. Marshall, Anal. Chim. Acta 142 (1982) 63.
[12] [13] [14] [15] [16] [17]
D.B. Marshall, C.H. Lochmuller, J. Chromatogr. Sci. 25 (1987) 262. L.C. Sander, S.A. Wise, Crit. Rev. Anal. Chem. 18 (1987) 299. J.J. Pesek, M.T. Matyska, J. Sep. Sci. 28 (2005) 1845. J.J. Pesek, M.T. Matyska, J. Liq. Chromatogr. Rel. Technol. 29 (2006) 1105. J.J. Pesek, M.T. Matyska, J.E. Sandoval, E. Williamsen, J. Liq. Chromatogr. Rel. Technol. 19 (1996) 2843. C.H. Chu, E. Jonsson, M. Auvinen, J.J. Pesek, J.E. Sandoval, Anal. Chem. 65 (1993) 808. J.E. Sandoval, J.J. Pesek, Anal. Chem. 63 (1991) 2634. J.J. Pesek, M.T. Matyska, M. Oliva, M. Evanchic, J. Chromatogr. A 818 (1998) 145. J.J. Pesek, M.T. Matyska, S. Prabhakaran, J. Sep. Sci. 28 (2005) 2437. J.J. Pesek, M.T. Matyska, LC/GC 24 (2006) 296. C. Yan, US Patent 5,453,763 (1995). J.J. Pesek, M.T. Matyska, unpublished results.