Anal. Chem. 2002, 74, 257-260
Ultrafast Sodium Dodecyl Sulfate Micellar Electrokinetic Chromatography with Very Acidic Running Buffers Miguel A. Rodrı´guez-Delgado,† Francisco J. Garcia-Montelongo,† and Alejandro Cifuentes*,‡
Department of Analytical Chemistry, University of La Laguna, 38071 Tenerife, Spain, and Department of Food Analysis, Institute of Industrial Fermentations (CSIC), Juan de la Cierva 3, 28006 Madrid, Spain
A new micellar electrokinetic chromatography (MEKC) procedure has been developed that allows the use of the anionic surfactant sodium dodecyl sulfate (SDS) together with separation buffers at pHs as low as 1. The technique is based on the employment of a high molecular weight polyethylenimine (PEI)-coated capillary that provides strong cathodal electroosmotic flows at acidic pHs when SDS is added to the running buffer. The procedure is easy to implement since the capillary coating is done by just flushing a PEI aqueous solution through the capillary and the subsequent steps are the same as those for any MEKC protocol. Moreover, the coating renewal provides reproducible separations between injections (migration time RSD values lower than 1.82 and 3.44% were obtained for the same day and three different days, respectively). The good possibilities of this system are demonstrated by showing the separation of a group of eight polyphenolic compounds within a separation time shorter than 2 min. This procedure allows one to extend the optimization of SDS-MEKC separations to the very acidic pH range too. Micellar electrokinetic chromatography1 (MEKC) using sodium dodecyl sulfate (SDS) as surfactant is by far the method most frequently used in capillary electrophoresis (CE) for the separation of neutral compounds.2 The extended use of SDS as surfactant in MEKC has been attributed to its low ultraviolet molar absorptivity, high aqueous solubility, low critical micelle concentration (cmc), low Krafft point, availability, and cost.2 Moreover, SDS has shown good selectivity for a large number of solute mixtures. These characteristics have made of SDS-MEKC the technique of choice for the separation of uncharged compounds. This procedure has been applied for the separation of, for example, antibiotics, vitamins, contaminants, drugs, and food constituents. However, one of the main limitations of SDS-MEKC is that the electroosmotic flow (EOF) must be strong enough in order to bear the SDS micelles toward the detection point. This limitation * Corresponding author: (e-mail)
[email protected]; (fax) 34-91-5644853. † University of La Laguna. ‡ Institute of Industrial Fermentations (CSIC). (1) Terabe, S.; Otsuka, K.; Ichikawa, K.; Tsuchiya, A.; Ando, T. Anal. Chem. 1979, 56, 111-113. (2) Khaledi, M. G. In High Performance Capillary Electrophoresis: Theory, Techniques and Applications; Khaledi, M. G., Ed.; John Wiley & Sons: New York, 1998; Chapter 3. 10.1021/ac010838k CCC: $22.00 Published on Web 11/30/2001
© 2002 American Chemical Society
can preclude in some cases the use of SDS-MEKC at low pHs where the EOF values are too weak as result of the lower ionization of the silanol groups onto the capillary wall.3 Although extreme acidic conditions limit the use of SDSMEKC, several authors4-7 have pointed out the advantages that could be attained by working with MEKC at the acidic pH range. Thus, modifications of pH of the SDS running buffer can bring about for ionizable and polar molecules variations of their capacity factor, what can be favorably used to achieve their separation.4 Accordingly, the use of low-pH solutions can result in higher resolutions and, under adequate conditions, in a wider elution range.5,7 Moreover, by making feasible the low-pH range, one of the most important advantages of MEKC, i.e., its flexibility by changing the chemical composition of the pseudostationary phases, can be enhanced.2 Also, it has been demonstrated that acidic pHs can be conveniently used to eliminate interferences from matrix during MEKC analysis of real samples.8 Despite these advantages, the number of works dealing with the use of SDS-MEKC at acidic conditions is scarce.5-11 This is mainly due, as mentioned above, to the limitation of the low EOF provided by bare capillaries at these pH values. To solve this limitation, two different approaches have been applied. The first one uses anionic polymer-coated capillaries, which provide EOF values independent of pH.6,7 Using these coated capillaries, it has been possible to obtain SDS-MEKC separations of neutral compounds at pH values as low as 3.15. However, to our knowledge, no SDS-MEKC separation at pHs lower than 3.15 has been shown with this type of column. More importantly, the coating procedure is laborious and time-consuming, and the capillary lifetime is only ∼60 h. The second approach is based on the fact that the mobility of SDS micelles is larger than EOF at acidic pHs in both bare silica and neutrally coated capillaries. Therefore, by injecting at (3) Poppe, H.; Cifuentes, A.; Kok, W. Th. Anal. Chem. 1996, 68, 888-893. (4) Otsuka, K.; Terabe, S.; Ando, T. J. Chromatogr. 1985, 348, 39-47. (5) Otsuka, K.; Terabe, S. J. Microcolumn Sep. 1989, 1, 150-154. (6) Landman, A.; Sun, P.; Hartwick, R. A. J. Chromatogr., A 1994, 669, 259262. (7) Sun, P.; Landman, A.; Barker, G. E.; Hartwick, R. A. J. Chromatogr., A 1994, 685, 303-312. (8) Cifuentes, A.; Bartolome´, B.; Go´mez-Cordove´s, C. Electrophoresis 2001, 22, 1561-1567. (9) Janini, G. M.; Muschik, G. M.; Issaq, H. J. J. Chromatogr., B 1996, 683, 29-35. (10) Quirino, J. P.; Terabe, S. Anal. Chem. 1998, 70, 1893-1901. (11) Quirino, J. P.; Terabe, S. Anal. Chem. 1999, 71, 1638-1644.
Analytical Chemistry, Vol. 74, No. 1, January 1, 2002 257
the negative end, it is possible to detect the analytes at the positive end.5,9-11 However, the neutral compounds migrating close to the EOF marker cannot be detected, too long analysis times are usually required, and its reproducibility using bare silica tubing is low. The goal of this work is, therefore, to show a new CE strategy that allows performing SDS-MEKC in an easy and reproducible manner using buffers of very low pH within short analysis times. EXPERIMENTAL SECTION Apparatus. The analyses were carried out in a P/ACE 2050 (Beckman Instruments, Fullerton, CA) CE apparatus, equipped with an UV-visible detector. The fused-silica capillary used was 27 cm total length (20 cm effective length) × 50 µm i.d. purchased from Composite Metal Services (Worcester, England). Injections were made at the anodic end using N2 pressure of 0.5 psi for 2 s (1 psi ) 6894.76 Pa). The instrument was controlled by a 486 DX2-66 MHz PC running the System GOLD software from Beckman. All measurements were carried out at 25 °C. Chemicals. All chemicals were of analytical reagent grade and used as received. SDS, trifluoroacetic acid, trichloroacetic acid, formic acid, and acetic acid from Merck (Darmstadt, Germany) were used for the CE running buffers at the different concentrations and pHs indicated below. Polyethylenimine (PEI) from Sigma (St. Louis, MO) was used for the capillary coating. Distilled water was deionized by using a Milli-Q gradient A10 system (Millipore, Bedford, MA). Polyphenolic standards were as follows: protocatechuic acid (1), (+)-catechin (2), 4-hydroxybenzoic acid (3), (-)-epigallocatechin (4), vanillic acid (5), caffeic acid (6), siringic acid (7), and p-coumaric acid (8) (all from Sigma, St. Louis, MO). Methanol, acetone, 2-propanol, and acetonitrile were HPLC grade from Merck (Darmstadt, Germany). All standards were dissolved in methanol at 1 mg/ml. EOF Measurements and MEKC Conditions. The running buffers used for the EOF calculation were as follows: 50 mM trifluoroacetic acid at pH 1; 50 mM trichloroacetic acid at pH 2; 50 mM formic acid at pH 3 and 4; and 50 mM acetic acid at pH 5. The SDS-containing buffers were as follows: 25 mM trifluoroacetic acid, 25 mM SDS at pH 1; 25 mM trichloroacetic acid, 25 mM SDS at pH 2; 25 mM formic acid, 25 mM SDS at pH 3 and 4; and 25 mM acetic acid, 25 mM SDS at pH 5. All the buffer pHs were adjusted as required by adding 0.1 M sodium hydroxide or 1 M hydrochloric acid and measured by a pH meter (40 pHmeter, Beckman). Running buffers used for the separation of polyphenolic compounds were from 10 to 50 mM SDS with and without different organic modifiers, containing 25 mM trifluoroacetic at pH 1.5 in all cases. Before first use, a new capillary was preconditioned by rinsing with 1 M NaOH for 15 min, followed by a 15-min rinse with deionized water. Between injections, the next washing protocol was applied (all the solutions were flushed by using a N2 pressure of 20 psi): 0.1 M sodium hydroxide for 0.5 min, water for 0.5 min, 5% PEI in water for 0.5 min, water for 0.5 min, and the separation electrolyte for 0.5 min. At the end of each day, the capillary was rinsed with deionized water for 2 min. Three injections (0.5 psi for 2 s) of a dissolution of acetone in water (5%, v/v) were carried out at the beginning of each day to stabilize 258 Analytical Chemistry, Vol. 74, No. 1, January 1, 2002
Figure 1. EOF versus pH for (9) 50 mM buffers at the indicated pHs and ([) 25 mM buffers plus 25 mM SDS at the indicated pHs. Separation conditions: fused-silica capillary with 50 µm of internal diameter, 27 cm of total length, and 20 cm effective length. Injection of acetone using a pressure of 0.5 psi for 1 s. Running voltage: 15 kV (when the 50 mM buffers were used together with the PEI-coated capillary -15 kV were applied). UV detection at 254 nm. (A) Bare silica capillary; (B) 5% PEI-coated capillary.
the whole system using a 25 mM trifluoroacetic, 40 mM SDS at pH 1.5 running buffer and 15 kV as separation voltage. RESULTS AND DISCUSSION In two previous works,12,13 we demonstrated the possibilities of using physically adsorbed high-molecular-mass PEI as coating for the separation of basic proteins and peptides by capillary electrophoresis. This coating is easily prepared, since it only requires flushing the PEI solution through the capillary, possesses good stability, and bears a high positive charge in a wide pH range that provides anodal EOFs with buffers of pH ranging from very acidic to basic. On the other hand, as demonstrated by Altria and Simpson14 and Tsuda,15 it is possible to reverse the EOF of a bare silica capillary from cathodal to anodal by adding to the separation buffer a cationic surfactant, such as, for example, cetyltrimethylammonium bromide (CTAB). According to all the above, it can be expected that the anodal EOF provided by a capillary bearing a high positive charge onto its inner wall (like capillaries coated with PEI), can be reversed by adding an anionic surfactant (like SDS) into the separation buffer, obtaining then an EOF toward the cathode. Moreover, the high number of positive charges onto the PEI coated-capillary wall is known to provide high anodal EOF values at very acidic pHs.12 Therefore, the addition of SDS to these acidic solutions should bring about a corresponding large number of surfactant anionic heads facing now the buffer and, as a consequence, giving rise to a strong cathodal EOF at very acidic (12) Erim, F. B.; Cifuentes, A.; Poppe, H.; Kraak, J. C. J. Chromatogr. 1995, 708, 356-361. (13) Cifuentes, A.; Poppe, H.; Kraak, J. C.; Erim, F. B. J. Chromatogr., B 1996, 681, 21-27. (14) Altria, K. D.; Simpson, C. F. Anal. Proc. 1986, 23, 453-454. (15) Tsuda, T. J. High. Resolut. Chromatogr. 1987, 10, 622-625.
Figure 2. Electrophoregrams of a set of polyphenolic compounds obtained with 25 mM trifluoroacetic acid at pH 1.50 and different SDS concentration: (A) 25 mM SDS; (B) 40 mM SDS. The capillary (total length 27 cm, effective length 20 cm, 50 µm i.d.) was coated with 5% PEI solution as indicated under Experimental Section. Run voltage 15 kV. Detection at 280 nm. Hydrodinamic injection at the anode for 2 s at 0.5 psi. For identification of compounds, see Experimental Section. Temperature 25 °C.
pHs. This strong EOF would allow carrying out fast SDS-MEKC separations even at very low pH values. To confirm this point, the next experiments were carried out. First, the EOF of a bare silica capillary was measured at pH values from 1 to 5 by injecting acetone in buffers with and without SDS. The results are shown in Figure 1A. As expected, by using these acidic buffers without SDS and a bare fused-silica capillary, low cathodal EOF values are obtained and they decrease with the pH until reaching values of practically zero at pHs of 2 and 1. To test the effect of SDS avoiding excessive heating generation, more dilute buffers were prepared (i.e., from 50 to 25 mM) containing 25 mM SDS. As can be seen in Figure 1A, stronger cathodal EOF are obtained with these SDS buffers compared with the EOFs obtained by using the 50 mM buffers without SDS. This is due to both the effect of the adsorption of ionized SDS on the inner surface wall, which increases the number of negative charges,16 and the lower ionic strength of the more dilute buffers containing SDS, as could be deduced from the lower electrical current that they provide under the same separation conditions (e.g., at pH 1 current values decreased from 92 to 63 µA; at pH 5 the variation (16) Terabe, S.; Utsumi, H.; Otsuka, K.; Ando, T.; Inomata, T.; Kuze, S.; Hanaoka, Y. J. High. Resolut. Chromatogr. 1986, 9, 666-670.
Figure 3. Electrophoregrams of a set of polyphenolic compounds obtained with 25 mM trifluoroacetic acid, 40 mM SDS at pH 1.50 with different organic modifiers in the separation buffer: (A) 10% methanol, (B) 5% 2-propanol, and (C) 5% acetonitrile. All the conditions as in Figure 2.
was from 20 to 16 µA). Under these conditions (i.e., bare silica capillary and acidic buffers with SDS), the mobility of SDS micelle determined with Sudan III is ∼2-fold larger than the EOF mobility at pH 3 and ∼5 times larger than the EOF mobility at pH 1, what precludes the use of SDS-MEKC at this pH range. Next, the capillary was coated with PEI, as described in the Experimental Section, and the EOF values at different pHs were measured by injecting acetone. Figure 1B shows the EOF values obtained using the PEI-coated capillary at pHs from 1 to 5 without SDS. The negative values mean that under these conditions the EOF is, as expected, toward the anode. The use of the buffers containing SDS brought about a reversal of the EOF direction, Analytical Chemistry, Vol. 74, No. 1, January 1, 2002
259
that is, the EOF is now toward the cathode, as can be deduced from the positive EOF values shown in Figure 1B. Moreover, these EOF values keep high even at the lowest pH tested, and their mobility in all the cases is stronger than the SDS micelle mobility, which indicates that under these conditions SDS-MEKC can be carried out without any constraint. To demonstrate this point, a buffer at pH 1.5 was arbitrarily chosen and a mixture of polyphenolic compounds was injected in the CE equipment, tuning the SDS concentration in order to obtain a suitable separation. Figure 2 shows the effect of SDS concentration onto the separation of a mixture of polyphenolic compounds. The effect of SDS concentration on separation is opposite to the effect normally obtained in MEKC. So, when 40 mM SDS was used, the migration times diminish with respect to 25 mM SDS. To confirm this point, a buffer with 10 mM SDS was tested and the same trend was observed; i.e., longer migration times than with 25 mM SDS were obtained. This could be justified by the equilibrium between SDS and the PEI layer. At higher SDS concentrations, the polymeric phase is more abundantly covered and the number of negative charges on the wall increases and with that the EOF value toward the cathode. Although the separation of Figure 2B using 40 mM SDS is relatively good, some polyphenolics comigrate, and therefore, a further optimization was necessary. To do this, different organic modifiers (i.e., methanol, 2-propanol, and acetonitrile) were tested to increase the resolution; the results obtained are shown in Figure 3. As can be seen in Figure 3A and C, the best results were achieved with 10% methanol and 5% acetonitrile, respectively. However, the use of 5% acetonitrile provides faster analyses, allowing the separation of the eight polyphenolics at pH 1.5 in less than 2 min. This better resolution can be explained through the decrease on electroosmotic flow17,18 (compare, for example, Figure 2B with the separations shown in Figure 3), together with (17) Lukkari, P.; Vuorela, H.; Riekkola, M. L. J. Chromatogr., A 1993, 655, 317324. (18) Schwer, C.; Kenndler, E. Anal. Chem. 1991, 63, 1801-1807. (19) Snopek, J.; Jelinek, I., Smolkova-Keulemansova, E. J. Chromatogr. 1988, 452, 571-579. (20) Cifuentes, A.; Bernal, J. L.; Diez-Masa, J. C. J. Chromatogr., A 1998, 824, 99-108.
260 Analytical Chemistry, Vol. 74, No. 1, January 1, 2002
Table 1. Reproducibility of Migration Times for Intraand Interday Analysis of a Mixture of Polyphenolic Compoundsa intraday (n ) 5)
interday (3 days, n ) 15)
peak no.
compound
t (min)
RSDb (%)
t (min)
RSDb (%)
1 2 3 4 5 6 7 8
protocatechuic acid (+)-catechin 4-hydroxybenzoic acid (-)-epigallocatechin vanillic acid caffeic acid siringic acid p-coumaric acid
1.27 1.34 1.38 1.44 1.49 1.56 1.70 1.92
0.22 0.55 0.46 0.75 0.89 1.19 1.04 1.82
1.28 1.35 1.39 1.45 1.49 1.57 1.71 1.94
0.88 0.74 0.70 0.97 0.92 1.22 2.04 3.44
aSeparation
condtions as In Figure 3C. b Relative standard deviation.
the changes induced on micellar size and retention factors of the phenolic compounds19,20 brought about by the addition of the organic solvent. Therefore, even though extremely acid buffers are used, separations can be optimized using the typical strategies utilized in MEKC. Table 1 shows the results for the reproducibility study carried out under the conditions of Figure 3C. As can be seen, relative standard deviations (RSD) of migration times were lower than 1.82% (n ) 5) for the same day, while these values increase up to 3.44% when the same experiment was repeated in three different days (n ) 15). A further optimization of the routine used to regenerate the capillary wall between injections should improve these reproducibility data. These experiments will be the subject of a forthcoming paper. ACKNOWLEDGMENT This work was supported by a University of La Laguna Precompetitive Research Project (243-60/99). Received for review July 24, 2001. Accepted September 30, 2001. AC010838K
