Mediated Separations by Capillary Zone Electrophoresis and Micellar

John G. Dorsey*. Department of Chemistry, Florida State University, Tallahassee, Florida 32306-3006. The addition of Ag(I) to the run buffer in capill...
11 downloads 0 Views 423KB Size
Download PDF
Articles Anal. Chem. 1996, 68, 415-424

Silver(I)-Mediated Separations by Capillary Zone Electrophoresis and Micellar Electrokinetic Chromatography: Argentation Electrophoresis Paul B. Wright

Department of Chemistry, University of Cincinnati, Cincinnati, Ohio 45221-0172 John G. Dorsey*

Department of Chemistry, Florida State University, Tallahassee, Florida 32306-3006

The addition of Ag(I) to the run buffer in capillary zone electrophoresis (CZE) and micellar electrokinetic chromatography (MEKC), containing sodium dodecyl sulfate, applies the principles of argentation chromatography to electrophoretic separations and is termed “argentation electrophoresis”. This technique is shown to provide a complementary method to CZE and MEKC for the separation of specific types of solutes that selectively complex with Ag(I). Baseline resolution in the CZE separation of nine sulfonamides is achieved by the addition of 50 mM silver nitrate to the run buffer. Retention mechanisms in MEKC separations can also be manipulated by the addition of Ag(I) to the micellar solution. Only slight resolution of a pair of sulfonamides was achieved under normal MEKC conditions. Upon the addition of Ag(I) to the mobile phase containing SDS micelles, baseline resolution of the compounds is shown. The retention order and resolution of five sulfonamides changed significantly when 25 mM Ag(I) was added to the SDScontaining buffer. The use of Ag(I) addition in MEKC is also applied to the separation of various other compounds that show selectivity toward Ag(I) complexation, including S-containing heterocycles and vitamin D compounds. The effect of the addition of Ag(I) on the elution range in MEKC is also investigated. A steady increase in the elution range is seen upon increasing the concentration of Ag(I). With a constant percentage of organic modifier (15%), the addition of higher concentrations of silver nitrate (2550 mM) results in elution ranges greater than 12. The results using Ag(I) as buffer additives in MEKC are also compared to studies utilizing a mixed counterion surfactant of sodium/silver dodecyl sulfate. Metal ions can complex with certain organic compounds such that various types of association products are formed. Many of the complexes are formed by the weak charge-transfer interaction of the compounds acting as electron donors and metal ions acting as electron acceptors. These weak complexes are normally 0003-2700/96/0368-0415$12.00/0

© 1996 American Chemical Society

unstable and exist in equilibrium with the solution components. Guha and Janak have reviewed the use of charge-transfer complexes of metals and their utility in chromatographic separations.1 The rates of formation for the complexes are generally very rapid, while the heats of formation are small, and data indicate that coordination forces are much smaller than those forces occurring when forming double bonds.1 Due to these properties, metal complexation of this type has been extensively exploited in both low- and high-performance liquid chromatography (HPLC). Although several metal ions can be used to invoke secondary chemical equilibria of this type and thereby alter the selectivity of a given chromatographic separation, the use of Ag(I) has been studied more extensively than that of other metal ions. The specific use of Ag(I) has been termed “argentation chromatography”. Argentation chromatographic separations have been successfully performed utilizing Ag(I) both as adsorbed or immobilized complexation sites on the stationary phase and as mobile phase additives (in the form of silver nitrate). Many types of compounds can complex with Ag(I) and have been successfully separated by argentation chromatography. Separations of particular interest are referenced here, but this does not represent an exhaustive survey of the argentation chromatography literature. Cooke et al. provide a general overview of the use of metal ions for selective separations in HPLC.2 Cagniant has published a more thorough discussion of the application of complexation phenomena in chromatography.3 Argentation chromatography is especially useful for N-, O-, and S-containing heterocyclics and isomers of various conjugated double-bond structures such as olefins in complex matrices. Vivilecchia et al. reported the separation of carcinogenic polynuclear azaheterocyclic hydrocarbons using a silver ion-impregnated stationary phase.4 Other studies utilizing Ag(I) immobilized on the stationary phase (1) Guha, O. K.; Janak, J. J. Chromatogr. 1972, 68, 325-343. (2) Cooke, N. H. C.; Viavattene, R. L.; Eksteen, R.; Wong, W. S.; Davies, G.; Karger, B. L. J. Chromatogr. 1978, 149, 391-415. (3) Cagniant, D. Chromatogr. Sci. 1992, 57, 149-195. (4) Vivilecchia, R.; Thiebaud, M.; Frei, R. W. J. Chromatogr. Sci. 1972, 10, 411416.

Analytical Chemistry, Vol. 68, No. 3, February 1, 1996 415

include the separation of stilbene derivatives5 and the evaluation of argentation chromatography for drug analysis.6 In addition, the degree of unsaturation in fatty acids was determined by argentation chromatography.7 Various other compounds have been separated by utilizing a mobile phase containing a low concentration of silver nitrate (typically 0.5%-2.5% by weight). Argentation chromatography has been useful in the separations of olefins and geometrical isomers,8,9 pharmaceutical compounds with varying double-bond structures such as vitamin D2 (ergocalciferol), vitamin D3 (cholecalciferol), and estrogenic compounds,10 and polyunsaturated hydrocarbons, fatty acid esters, triglycerides, and various N-, O-, and S-containing heterocyclic hydrocarbons.2,9 Even though the use of argentation chromatography was more prevalent in the mid-to-late 1970s, the specificity of Ag(I) complexation still provides an effective means of effecting the separation of certain types of compounds, as shown by recent publications utilizing argentation chromatography. The technique has been used recently for the efficient separation of polynuclear aromatic hydrocarbons;11 olefins, heteroatomic compounds, and lipids;3,12 triglycerides;13,14 geometrical isomers of linolenic acid;15 and fatty acid esters and methyl esters.16,17 This paper reports the first application of silver ion complexation in capillary zone electrophoresis (CZE) and micellar electrokinetic chromatography (MEKC) and is termed “argentation electrophoresis”. CZE has developed over the past decade into an electrophoretic separation method capable of very fast and efficient separations of charged solutes.18-23 The use of narrowbore capillaries, i.e., 25-75 µm i.d. or smaller, allows the application of fairly large electric field strengths without the problem of band broadening caused by excessive Joule heating.20 A major drawback of CZE, however, is its inherent inability to separate neutral compounds. Terabe and co-workers were the first to propose the use of surfactant buffer additives for capillary electrophoretic separations.24,25 This mode of CE, also designated micellar electrokinetic (5) Fuggerth, E. J. Chromatogr. 1979, 169, 469-473. (6) Takahagi, H. Bunseki 1978, 11, 810-815. (7) Huq, M. S.; Khan, M. S.; Rubbi, S. F. Bangladesh J. Sci. Ind. Res. 1979, 14, 159-169. (8) Schomburg, G.; Zegarski, K. J. Chromatogr. 1975, 114, 174-178. (9) Vonach, B.; Schomburg, G. J. Chromatogr. 1978, 149, 417-430. (10) Tscherne, R. J.; Capitano, G. J. Chromatogr. 1977, 136, 337-341. (11) Dunn, J. A.; Holland, K. B.; Jezorek, J. R. J. Chromatogr. 1987, 394, 375381. (12) Aitzetmueller, K.; Goncalves, G.; Lireny, A. J. Chromatogr. 1990, 519, 349358. (13) Neff, W. E.; Adlof, R. O.; List, G. R.; El-Agaimy, M. J. Liq. Chromatogr. 1994, 17, 3951-3969. (14) Kemper, K.; Melchert, H. U.; Rubach, K. Lebensmittelchem. Gerichtl. Chem. 1988, 42, 105-108. (15) Juaneda, P.; Sebedio, J. L.; Christie, W. W. J. High Resolut. Chromatogr. 1994, 17, 321-324. (16) Nikolova-Damyanova, B.; Christie, W. W.; Herslof, B. J. Chromatogr. A 1995, 693, 235-239. (17) Ulberth, F.; Achs, E. J. Chromatogr. 1990, 504, 202-206. (18) Ewing, A. G.; Wallingford, R. A.; Olefirowicz, T. M. Anal. Chem. 1989, 61, 292A-303A. (19) Jorgensen, J. W.; Lukacs, K. D. J. Chromatogr. 1981, 218, 209-216. (20) Jorgenson, J. W.; Lukacs, K. D. Anal. Chem. 1981, 53, 1298-1302. (21) Jorgenson, J. W.; Lukacs, K. D. Science 1983, 222, 266-272. (22) Grossman, P. D.; Colburn, J. C. Capillary Electrophoresis Theory and Practice; Academic Press: San Diego, CA, 1992; 352 pp. (23) Weinberger, R. Practical Capillary Electrophoresis; Academic Press Inc.: San Diego, CA, 1993; 312 pp. (24) Terabe, S.; Otsuka, K.; Ichikawa, K.; Tsuchiya, A.; Ando, T. Anal. Chem. 1984, 56, 111-113.

416

Analytical Chemistry, Vol. 68, No. 3, February 1, 1996

capillary chromatography (MECC),26 provides high efficiency and has the advantage of providing a form of secondary chemical equilibria for the separation of neutral species. At surfactant concentrations above the critical micelle concentration, micelles are dynamically formed in the carrier electrolyte, and separation is based on differential chromatographic partitioning of the solutes between the aqueous (buffer) phase and the micellar phase. The technique has proven to be very useful for the separation of a wide variety of compounds including both charged and uncharged species and continues to be at the forefront of separation science research. Some previous studies have made use of complexation phenomena in CZE and MEKC. Metal ions have been determined by CZE using a wide variety of complexing reagents added to the separation buffer.27-34 MEKC separations of metal chelates have also been described.35 Other studies have utilized metal ions or other charged species as complexing agents to separate various compounds.36 Walbroehl and Jorgenson showed the electrophoretic separation of various organic molecules by employing weak complexation with tetraalkylammonium ion.37 Similarly, the tetraalkylammonium ion was employed in MEKC and shown to improve the resolution of various ionic substances.38 Chiral separations of amino acid enantiomers were achieved by the addition of cupric ions and a chiral selector to the buffer in CZE.39,40 Owing to the variety of N-, O-, and S-containing sites on peptides and proteins, the effect of metal ion addition, including Cu(II), Zn(II), Ca(II), and Fe(II), on the CZE separations of peptides and proteins has also been investigated. Mosher studied the effect of Cu(II) and Zn(II) buffer additives on the separation of dipeptides containing histidine.41 Polycarboxylic acids were determined by CZE utilizing copper complexation.42 Cai and El Rassi developed capillaries with surface-bound iminodiacetic acidZn chelating moities for the separation of dilute protein samples.43 CZE and MEKC separations of proteins can also be affected by the presence of Ca(II) and Zn(II).44 The addition of Mg(II) to SDS-containing buffers was also shown to be an effective way to enhance the selectivity of MEKC (25) Terabe, S.; Otsuka, K.; Ando, T. Anal. Chem. 1985, 57, 834-841. (26) Burton, D. E.; Sepaniak, M. J.; Maskarinec, M. P. J. Chromatogr. Sci. 1986, 24, 347-350. (27) Morin, P.; Francois, C.; Dreux, M. J. Liq. Chromatogr. 1994, 17, 38693888. (28) Regan, F. B.; Meaney, M. P.; Lunte, S. M. J. Chromatogr. B 1994, 657, 409-417. (29) Motomizu, S.; Oshima, M.; Kuwabara, M.; Obata, Y. Analyst 1994, 119, 1787-1792. (30) Vogt, C.; Conradi, S. Anal. Chim. Acta 1994, 294, 145-153. (31) Timerbaev, A. R.; Semenova, O. P.; Bonn, G. K.; Fritz, J. S. Anal. Chim. Acta 1994, 296, 119-128. (32) Shi, Y.; Fritz, J. S. J. Chromatogr. 1993, 640, 473-479. (33) Buchberger, W.; Semenova, O. P.; Timerbaev, A. R. J. High Resolut. Chromatogr. 1993, 16, 153-156. (34) Swaile, D. F.; Sepaniak, M. J. Anal. Chem. 1991, 63, 179-184. (35) Saitoh, T.; Hoshino, H.; Yotsuyanagi, T. J. Chromatogr. 1989, 469, 175181. (36) Snopek, J.; Jelinek, I.; Smolkova-Keulemansova, E. J. Chromatogr. 1988, 452, 571-590. (37) Walbroehl, Y.; Jorgenson, J. W. Anal. Chem. 1986, 58, 479-481. (38) Nishi, H.; Tsumagari, N.; Terabe, S. Anal. Chem. 1989, 61, 2434-2439. (39) Gassman, E.; Kuo, J. E.; Zare, R. N. Science 1985, 230, 813-814. (40) Gozel, P.; Gassmann, E.; Michelson, H.; Zare, R. N. Anal. Chem. 1987, 59, 44-49. (41) Mosher, R. A. Electrophoresis 1990, 11, 765-769. (42) Wiley, J. P. J. Chromatogr. A 1995, 692, 267-274. (43) Cai, J.; El Rassi, Z. J. Liq. Chromatogr. 1993, 16, 2007-2024. (44) Kajiwara, H. J. Chromatogr. 1991, 559, 345-356.

separations of oligosaccharides.45 Acetylacetonato complexes with Cr(III), Co(III), Rh(III), Pt(II), and Pd(II) were also utilized in MEKC.46 Cohen and co-workers observed enhanced selectivity in the MEKC separation of bases, nucleosides, and nucleotides when Cu(II), Zn(II), and Mg(II) were added to the buffer containing SDS micelles.47 These metals are known to be electrostatically attracted to the negatively charged SDS micelle surface.48 Selectivity enhancement of these solutes was attributed to differential complexation with the metal ions added to the SDScontaining buffer.47 In addition to the ability to manipulate chromatographic selectivity, the use of metal ions as buffer additives in MEKC provides other possible advantages, including minimized band broadening due to rapid kinetics of the association-dissociation process and elution range, or elution window, extension.47 The limited elution range continues to be an obstacle to performing complex analyses by MEKC. Much research has been dedicated to increasing the ratio tmic/t0, defined as the elution range, and thereby increasing the number of components capable of being separated by the technique.25,49-61 The aim of this work is to apply the principles of argentation chromatography to separations using capillary electrophoresis and micellar electrokinetic chromatography. The use of Ag(I) as a mobile phase additive in CZE and MEKC was investigated for separations of specific types of solutes that can form complexes with Ag(I). This study shows that invoking secondary equilibria in ways other than the use of micelles can be advantageous and provide a complementary method to conventional MEKC for the separation of certain types of solutes, i.e., those containing π-electrons and compounds with S-, O-, and N-containing lone electron pairs. Compounds such as these can be difficult to separate by conventional CZE or MEKC. EXPERIMENTAL SECTION Apparatus. Capillary zone electrophoresis and micellar electrokinetic chromatography were performed using a Beckman (Fullerton, CA) P/ACE system 2100 equipped with an uncoated fused silica capillary, obtained from Polymicro Technologies Inc. (Phoenix, AZ), with an inner diameter of 50 µm and an outer diameter of 365 µm. Capillary length was varied in 10 cm (45) Taverna, M.; Baillet, A.; Baylocq-Ferrier, D. Chromatographia 1993, 37, 415-422. (46) Saitoh, K.; Kiyohara, C.; Suzuki, N. J. High Resolut. Chromatogr. 1991, 14, 245-248. (47) Cohen, A. S.; Terabe, S.; Smith, J. A.; Karger, B. L. Anal. Chem. 1987, 59, 1021-1027. (48) Oko, M. U.; Venable, R. L. J. Colloid Interface Sci. 1971, 35, 53-59. (49) Balchunas, A. T.; Sepaniak, M. J. Anal. Chem. 1987, 59, 1466-1470. (50) Balchunas, A. T.; Swaile, D. F.; Powell, A. C.; Sepaniak, M. J. Sep. Sci. Technol. 1988, 23, 1891-1904. (51) Balchunas, A. T.; Sepaniak, M. J. Anal. Chem. 1988, 60, 617-621. (52) Gorse, J.; Balchunas, A. T.; Swaile, D. F.; Sepaniak, M. J. J. High Resolut. Chromatogr. Chromatogr. Commun. 1988, 11, 554-559. (53) Foley, J. P. Anal. Chem. 1990, 62, 1302-1308. (54) Rasmussen, H. T.; Goebel, L. K.; McNair, H. M. J. Chromatogr. 1990, 517, 549-555. (55) Terabe, S.; Ishihama, Y.; Nishi, H.; Fukuyama, T.; Otsuka, K. J. Chromatogr. 1991, 545, 259-368. (56) Vindevogel, J.; Sandra, P. Anal. Chem. 1991, 63, 1530-1536. (57) Cole, R. O.; Sepaniak, M. J. LC-GC 1992, 10, 380-385. (58) Kaneta, T.; Tanaka, S.; Taga, M.; Yoshida, H. J. Chromatogr. 1992, 609, 369-374. (59) Cai, J.; El Rassi, Z. J. Chromatogr. 1992, 608, 31-45. (60) Tsai, P.; Patel, B.; Lee, C. S. Anal. Chem. 1993, 65, 1439-1442. (61) Ahuja, E. S.; Little, E. L.; Nielson, K. R.; Foley, J. P. Anal. Chem. 1995, 67, 26-33.

increments from 37 to 77 cm, with an effective length of 30-70 cm, respectively. The temperature was controlled via liquid cooling at 26 °C unless otherwise indicated. The possibility of Joule heating was minimized by the use of biological buffers, which carry lower currents than phosphate or borate buffers typically employed in CZE and MEKC. Currents did not exceed 50 µA (using a 47 cm column) when the biological buffers were used. Slightly higher currents resulted from the use of either an acetate buffer or a shorter capillary, while lower currents were obtained with longer capillaries. Samples were introduced by hydrodynamic injection for 3-5 s. Normal polarity separations were performed with on-line detection near the cathode at either 254 or 280 nm. Electropherograms and chromatograms were obtained through Beckman P/ACE software and/or System Gold software (Beckman). Reagents. All samples and buffers were prepared with analytical reagent grade chemicals and ∼18 MΩ deionized water filtered through a Barnstead Nanopure system (Dubuque, IA). Analytical grade sodium dodecyl sulfate (SDS), 3-cyclohexylamino1-propanesulfonic acid (CAPS) buffer, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES) buffer, ammonium acetate, silver nitrate, sulfonamides, ergocalciferol, and cholecalciferol were purchased from Sigma Chemicals (St. Louis, MO). Acetic acid and organic modifiers such as methanol, acetonitrile, and 1-propanol were obtained from Fisher Scientific (Fair Lawn, NJ). Thioxanthone and dibenzothiophene were obtained from Fluka (Ronkonkoma, NY). Surfactant solutions at a concentration well above the critical micelle concentration were prepared by dissolving SDS in 30 mM buffer solutions. The buffer pH was adjusted with either 100 mM acetic acid or HCl (Fisher Scientific), or 1 M ammonium hydroxide (Sigma Chemicals). For CZE and MEKC experiments utilizing Ag(I) complexation, silver nitrate was added prior to pH adjustment. All solutions were filtered through a 0.45 µm nylon Acrodisk 13 filter (Gelman Sciences, Ann Arbor, MI) before CZE or MEKC. Procedure. Before each run, the capillary was rinsed using high pressure for 2 min with 0.1 M ammonium hydroxide, followed by 2-3 min of rinsing with deionized water, then 2-3 min with buffer prior to injection. Special attention was paid to assure sufficient regeneration and equilibration of the capillary to promote reproducibility of the capillary surface and, hence, electroosmotic flow. When Ag(I) is added to the buffer, extra care needs to be taken in terms of rinsing in order to prevent the possible formation of hydroxide precipitates. The use of ammonium hydroxide as the rinsing base rather than sodium hydroxide helps prevent precipitation in the case of the silver ion. It should be noted, however, that this may not be the case for other metal ions such as Mg2+, for example. We also tested for the photolytic conversion of Ag(I) to Ag(0) within the capillary and found no detection problems. The detection window that was burned remained transparent despite the use of high concentrations of silver nitrate in the run buffer. In MEKC experiments, the parameters t0 and tmic (defined as the retention times obtained for the electroosmotic flow and micelles, respectively) were experimentally measured. The migration of an uncharged and uncomplexed species, t0, is given by the solvent peak or injection of a small amount of methanol and was measured by the resulting peak. The migration of the micelle, tmic, was measured by the addition of either Sudan III or decanophenone to the micellar solution. There is no indication that either of these two comAnalytical Chemistry, Vol. 68, No. 3, February 1, 1996

417

Figure 1. CZE separation of sulfaguanidine and sulfanilamide as a function of Ag(I) concentration: (A) 0, (B) 5, (C) 15, and (D) 25 mM silver nitrate. pH 7 HEPES buffer, no organic modifier; 47 cm capillary, 40 cm effective length; 10 kV.

pounds interacts with Ag(I). These two compounds are assumed to be fully partitioned or completely solubilized by the micelle and therefore provide markers for it. RESULTS AND DISCUSSION Effect of Ag(I) on CZE Separations. Given the established interaction of Ag(I) with various compounds taken from the argentation chromatography literature, especially N-, O-, and S-containing heterocyclics, selected sulfonamides were chosen as test compounds to evaluate the use of Ag(I) as an additive in CZE. These compounds have widely varying structures, hydrophobicities, and pKa values, which make them versatile test solutes. Most CZE studies were conducted using a pH 4 acetate buffer such that all of the sulfonamides studied were electrically neutral species. Some experiments, however, utilized a pH 7 HEPES buffer. These experiments were limited to only a few sulfonamides that would be neutral at this pH. Since it is expected that silver ions would be attracted to the negatively charged walls of the capillary, a separation of sulfanilamide and sulfaguanidine with a pH 7 HEPES buffer (both solutes are neutral at this pH) was performed in order to study both the resolution of the two compounds and the efficiency as a function of increasing Ag(I) concentration. Figure 1 shows the electropherograms obtained as Ag(I) concentration is increased (0-25 mM silver nitrate). As expected, under normal CZE conditions where there is no Ag(I) added to the buffer, the two neutral compounds coelute (Figure 1A). As the concentration of silver nitrate is gradually increased, the resolution of the two compounds improves. At Ag(I) concentrations over 15 mM, the two sulfonamides are baseline resolved. For this particular separation, increasing the silver ion concentration above 25 mM (e.g., 50 mM) did not show much further improvement in the resolution of the compounds, yet increased the analysis time. Sulfaguanidine elutes before sulfa418

Analytical Chemistry, Vol. 68, No. 3, February 1, 1996

nilamide owing to its more favorable interaction with Ag(I). Sulfaguanidine contains a double bond as well as nitrogen atoms capable of forming weak complexes with Ag(I), thus creating a positively charged species, which causes faster migration and separation from sulfanilamide, which is not expected to interact with Ag(I).2 MEKC results herein also indicate that sulfanilamide does not interact with Ag(I). It should be noted that minor variations in peak heights (and peak height ratios) with increasing Ag(I) concentration is merely caused by variations in the relative concentrations of the solutes since samples were prepared just prior to performing the separation at each concentration. Efficiencies were estimated using the exponentially modified Gaussian approach of Foley and Dorsey.62 Table 1 shows the data for efficiency as a function of silver nitrate concentration for the separation of sulfaguanidine and sulfanilamide, under the same conditions as in Figure 1. In examining Figure 1, it is obvious that resolution increases dramatically in the CZE separation of sulfanilamide and sulfaguanidine as Ag(I) concentration is increased. However, the data in Table 1 show that although the selectivity improves with the addition of Ag(I) to the CZE buffer, efficiencies are somewhat degraded. The loss of efficiency can be attributed to interactions of the solutes with silver ions that are electrostatically attracted to the negatively charged wall of the capillary. It should also be noted that while no systematic study of migration time reproducibility was investigated, migration times show an apparent increase in run-to-run variability as the Ag(I) concentration increased. It is well known that the addition of electrolyte causes a decrease in electroosmotic flow velocity and thereby affects the migration time of solutes separated by CE techniques.22,23,45,47,63-67 Also (62) Foley, J. P.; Dorsey, J. G. Anal. Chem. 1983, 55, 730-737. (63) Issaq, H. J.; Atamna, I. Z.; Muschik, G. M.; Janini, G. M. Chromatographia 1991, 32, 155-161. (64) Altria, K. D.; Simpson, C. F. Chromatographia 1987, 24, 527-532.

Table 1. Efficiency, Electroosmotic Mobility, and Electrophoretic Mobility Data for Sulfaguanidine and Sulfanilamide as a Function of Silver Nitrate Concentration systema

solute

migration timeb (min ( SD)c

CZE CZE CZE + 5 mM AgNO3 CZE + 5 mM AgNO3 CZE + 15 mM AgNO3 CZE + 15 mM AgNO3 CZE + 25 mM AgNO3 CZE + 25 mM AgNO3

S-guan/S-anil S-anil S-guan S-anil S-guan S-anil S-guan S-anil

4.34 ( 0.10 4.24 ( 0.10f 5.75 ( 0.05f 5.81 ( 0.05f 8.32 ( 0.06 8.66 ( 0.06 9.94 ( 0.24 10.94 ( 0.27

µeo (×10-4)d (cm2 V-1 s-1)

µep (×10-5)e (cm2 V-1 s-1) 0.00

7.39 0.593 5.39 1.46 3.62 2.92 2.86

N 80 000 120 000 90 000 79 000 37 000 18 000 27 000 4 000

a Data obtained using a HEPES buffer (30 mM), pH 7, 47 cm capillary (40 cm effective length), 10 kV. b Triplicate runs were done in order to obtain average migration times. c Errors represent standard deviations. d Electroosmotic mobilities calculated from the migration time for sulfanilamide. e Electrophoretic mobility of sulfanilamide is assumed to be zero. Electrophoretic mobilities in this column are calculated for sulfaguanidine. f These data were obtained through an independent experiment where only one analyte was present.

included in Table 1, therefore, are the calculated electroosmotic and electrophoretic mobilities for the separations in Figure 1. These values should present more clearly the effect of Ag(I) addition on the separation. They are calculated using the following equations:

µeo ) (lL)/(t0V)

(1)

µapp ) (lL)/(tmV)

(2)

µapp ) µep + µeo

(3)

where µeo is the electroosmotic mobility, µapp is the apparent mobility or overall observed mobility of the analyte, µep is the electrophoretic mobility of the analyte, t0 is the migration time of an uncharged, uncomplexed compound, tm is the migration time of the analyte, l is the effective capillary length to the detector, L is the total capillary length, and V is the applied voltage. As the Ag(I) concentration increases, and hence the ionic strength increases, the electroosmotic mobility decreases. As a result, overall migration times are longer. However, as Ag(I) concentration increases, the electrophoretic mobility of sulfaguanidine increases owing to the positive charge of the weak complex formed. The Ag(I)-analyte interaction, therefore, is the major equilibrium process governing this separation. To demonstrate the ability of the technique to separate more complex samples, the separation of nine sulfonamides was performed as a function of silver ion concentration. The resulting electropherograms for the separation of sulfadimethoxine, sulfaguanidine, sulfamerazine, sulfaquinoxaline, sulfadiazine, sulfathiazole, sulfanilamide, sulfacetamide, and sulfabenzamide are presented in Figure 2. Without Ag(I) added to the buffer, Figure 2A, most of the sulfonamides coelute as neutral species when a pH 4 acetate buffer is used. The presence of two peaks eluting after the main peak represents sulfacetamide and sulfabenzamide, which may be present in some dissociated negatively charged form, due to their low pKa values. Sulfanilamide, therefore, is regarded as a true marker of electroosmotic flow in these studies since it is neutral at this pH and does not interact with Ag(I). The pH measured is the apparent pH, since this buffer system (65) Bruin, G. J. M.; Chang, J. P.; Kuhlman, R. H.; Zegers, K.; Kraak, J. C.; Poppe, H. J. Chromatogr. 1989, 471, 429. (66) Nashabeh, W.; El-Rassi, Z. J. Chromatogr. 1990, 514, 57. (67) Zhang, Y.; Warner, I. M. J. Chromatogr. A 1994, 688, 293-300.

Figure 2. CZE separation of nine sulfonamides as a function of Ag(I) concentration: (A) 0 mM silver nitrate (47 cm capillary), (B) 15 mM silver nitrate (37 cm capillary), and (C) 80 mM silver nitrate (47 cm capillary). pH 4 acetate buffer; 15% ACN organic modifier; 20 kV. Peaks: 1, sulfadimethoxine; 2, sulfaguanidine; 3, sulfamerazine; 4, sulfaquinoxaline; 5, sulfadiazine; 6, sulfathiazole; 7, sulfanilamide; 8, sulfacetamide; and 9, sulfabenzamide.

Analytical Chemistry, Vol. 68, No. 3, February 1, 1996

419

also contained 15% acetonitrile (ACN) to aid in the solubility of the more hydrophobic sulfonamides. Since pKa values can vary with added organic modifier, the presence of more than one peak under these CZE conditions is not unexpected. In general, however, the neutral sulfonamides coelute as expected under typical CZE conditions (Figure 2A). Upon increasing the concentration of Ag(I), the resolution of the sulfonamides improves greatly. It should be noted that the detection of both sulfadimethoxine and sulfaquinoxaline was difficult due to the low solubility of these sulfonamides, even when a small percentage of ACN was added to the buffer. The separation improves upon the addition of 15 mM Ag(I) (Figure 2B), while the addition of 80 mM silver nitrate results in baseline resolution of all nine compounds, with the exception of sulfamerazine and sulfaquinoxaline (Figure 2C). Again, the coelution of these two compounds is assumed, but the poor solubility of sulfaquinoxaline precludes detection of this analyte. An independent injection of sulfaquinoxaline showed that it migrated with or near that of sulfamerazine. It should also be noted that these two compounds are partially resolved under 50 mM silver ion conditions, supporting the notion that sulfaquinoxaline was actually not detected under 80 mM silver ion conditions due to poor solubility. The calculated electroosmotic and electrophoretic mobilities as a function of Ag(I) concentration for the separations in Figure 2 are shown graphically in Figure 3. As expected, the electroosmotic mobility decreases as the Ag(I) concentration increases due to the increase in ionic strength (Figure 3A). Individual electrophoretic mobilities of the neutral sulfonamides are shown to increase as the concentration of Ag(I) increases, and the general trends are evident in Figure 3B. This again shows that the Ag(I)-analyte interaction is the main mechanism providing the separation of the neutral sulfonamides. The negative compound, sulfabenzamide, shows anomalous behavior with increasing Ag(I) concentration. The relationship is nonlinear and goes through a maximum in the concentration range studied. As predicted from their structures, the electrophoretic mobilities of sulfanilamide, taken here as the marker of electroosmotic flow, and sulfacetamide do not change much as a function of Ag(I) concentration due to little or no interaction with the metal ion. Effect of Ag(I) on MEKC Separations. In addition to partitioning mechanisms, the addition of metal ions such as Ag(I) can alter selectivity in MEKC separations through complexation phenomena. Using a pH 7 HEPES buffer containing SDS, the electrokinetic separation of sulfanilamide and sulfaguanidine was studied as a function of silver ion concentration. At this pH, both solutes are neutral. As shown in Figure 4A, under normal MEKC conditions using an SDS concentration of 25 mM and an applied voltage of 10 kV, the two compounds are not baseline resolved. By employing Ag(I) complexation as an additional form of secondary chemical equilibria, the two compounds become well resolved as the concentration of Ag(I) in the buffer is increased, as shown in Figure 4B-D, with sulfanilamide eluting first, followed by sulfaguanidine. The greatest improvement in resolution is seen upon increasing the concentration of Ag(I) to 15 mM (Figure 4C). A further increase in metal ion concentration resolves the solutes more than necessary for an adequate separation and only serves to increase the analysis time, but dramatically illustrates the capabilities of the technique for selectivity manipulation. 420 Analytical Chemistry, Vol. 68, No. 3, February 1, 1996

Figure 3. Electroosmotic and electrophoretic mobilities as a function of Ag(I) concentration in the separation of nine sulfonamides: (A) µeo vs Ag(I) concentration and (B) µep vs Ag(I) concentration for nine sulfonamides. pH 4 acetate buffer; 15% ACN organic modifier; 20 kV.

The exact retention mechanism is not known, but an equilibrium exists between Ag+ ions electrostatically attracted to the micelle and those in solution. Therefore, the interaction of solutes with Ag(I) is expected to be a function of these same equilibrium processes. The data here suggest that for this particular separation of two neutral and relatively polar compounds, elution order corresponds to the degree of association of the solute with Ag(I) ions electrostatically attracted to the micelle surface. Due to the double-bond character of this solute, as well as its lone pair electrons, sulfaguanidine would be expected to interact more with Ag(I) (present at the micelle surface) and therefore elutes after sulfanilamide. In addition, electroosmotic mobility, determined from the retention time of an unretained compound (MeOH), was calculated as a function of Ag(I) concentration. µeo decreased by more than 35% on going from 0 to 25 mM Ag(I). The decrease in µeo, caused by an increase in Ag(I) concentration, results in longer analysis time as the apparent mobility of the micelle decreases. The separation, however, of these two neutrals is improved due to the selective interaction with Ag(I). Table 2 shows the effect that the addition of Ag(I) has on the retention data for these two compounds. The data in the table

Figure 4. MEKC separation of sulfanilamide and sulfaguanidine as a function of Ag(I) concentration: (A) 0, (B) 5, (C) 15, and (D) 25 mM silver nitrate. pH 7 HEPES buffer; 25 mM SDS; no organic modifier; 47 cm capillary, 40 cm effective length; 10 kV.

Table 2. Selectivity, k′, and Resolution Data for Sulfanilamide and Sulfaguanidine as a Function of Silver Nitrate Concentration system

solute

k′

MEKC MEKC MEKC + 5 mM Ag(I) MEKC + 5 mM Ag(I) MEKC + 15 mM Ag(I) MEKC + 15 mM Ag(I) MEKC + 25 mM Ag(I) MEKC + 25 mM Ag(I)

S-anil S-guan S-anil S-guan S-anil S-guan S-anil S-guan

0.11 0.19 0.09 0.19 0.13 0.53 0.14 1.41

R 1.7 2.1 4.1 10.1

are calculated from the following chromatographic equations, adapted for MEKC:

k′ ) (tR - t0)/t0[1 - (tR/tmic)]

(4)

R ) k2′/k1′

(5)

The values for t0 and tmic were determined experimentally as described earlier, and tR is the measured retention time of the solute. k′ is the capacity factor, given by eq 4, and R is the selectivity, given by eq 5. As apparent from the chromatograms, selectivity and resolution are greatly improved by the addition of Ag(I) to the surfactant-containing buffer. Under normal MEKC conditions, where no silver nitrate is added, k′ values for both solutes are low, indicating relatively little partitioning of the solutes to the micelle (Figure 4A). Upon increasing the Ag(I) concentration, k′ values for sulfanilamide show little change, which indicates

minimal interaction with the micelle or silver ions at the surface of the micelle. On the other hand, k′ values for sulfaguanidine increase greatly as the silver ion concentration reaches 15-25 mM. The change in k′ for sulfaguanidine not only demonstrates the effect of the addition of Ag(I) on retention but also provides support for the presence of silver ions at or near the micelle surface. These data show that the addition of relatively small concentrations of Ag(I) to typical anionic surfactant systems can dramatically affect MEKC separations and the governing retention characteristics of the micellar system. The presence of silver ions at or near the micellar surface provides complexation sites for solutes capable of interacting with Ag(I), thus providing an additional way to manipulate selectivity in MEKC. The use of Ag(I) as a buffer additive in MEKC is applied to the separation of various types of compounds that have shown selectivity toward Ag(I) complexation. The separations of sulfonamides, two S-containing heterocyclics, and vitamins D2 and D3 are shown. Figure 5 shows the separation of five sulfonamides without and with 25 mM silver nitrate added to the MEKC buffer. These particular sulfonamides were chosen on the basis of varying structure, charge, and propensity for silver ion complexation. Using MEKC alone, the resolution of the five sulfonamides is incomplete, as the first two compounds (sulfanilamide and sulfaguanidine) are not quite baseline resolved, and the other three compounds (sulfacetamide, sulfapyridine, and sulfamerazine) coelute (Figure 5A). With the addition of 25 mM silver nitrate, the separation of the five compounds is greatly improved, while the elution order changes dramatically. This indicates a change in the retention mechanisms governing the MEKC separation. The two compounds which elute first and second (sulfanilamide and sulfacetamide) would not be expected to Analytical Chemistry, Vol. 68, No. 3, February 1, 1996

421

Figure 5. MEKC separation of five sulfonamides: (A) 0 and (B) 25 mM silver nitrate added. pH 7 HEPES buffer; 25 mM SDS; no organic modifier; 47 cm capillary, 40 cm effective length; 20 kV. Peaks: 1, sulfanilamide; 2, sulfaguanidine; 3, sulfacetamide; 4, sulfapyridine; and 5, sulfamerazine.

interact much with either Ag(I) or the micelle, owing to their structures and/or charge at pH 7. The other three compounds would be expected to interact more with Ag(I) and the micelle and therefore to be retained longer. Sulfanilamide and sulfaguanidine, which were not completely separated by MEKC, became very well resolved when silver ions were added to the MEKC buffer. The other three compounds that coeluted under simple MEKC conditions are baseline resolved with the addition of 25 mM silver nitrate to the buffer. Unfortunately, sulfaguanidine and sulfapyridine show similar selectivity toward Ag(I) complexation and coelute at this concentration. The MEKC separations of a pair of S-containing heterocyclics are shown in Figure 6. These solutes are more hydrophobic than the sulfonamides and coelute near the micelle (k′ ) 52 for the two solutes in the absence of Ag(I)), unlike the sulfonamides, which eluted early in the absence of Ag(I). With the addition of 25 mM silver nitrate, thioxanthone (peak 1) and dibenzothiophene (peak 2) are nearly baseline resolved, yet still elute close to the micelle. The k′ value for thioxanthone was 14.7, while k′ ) 22.7 for dibenzothiophene. Again, the retention mechanism for this separation process is not well understood, but the retention data support a shift from a separation based primarily on hydrophobicity to one based on metal ion complexation. Unlike the more polar sulfonamides, these compounds are less retained (k′ values decrease) upon the addition of Ag(I) to the surfactant-containing buffer. This result suggests a slightly different separation mechanism that is a function of the type of compounds being separated, e.g., structural differences and differences in hydrophobicities. Finally, using a CAPS buffer at pH 10, ergocalciferol (vitamin D2) and cholecalciferol (vitamin D3) were effectively separated 422 Analytical Chemistry, Vol. 68, No. 3, February 1, 1996

Figure 6. MEKC separation of dibenzothiophene and thioxanthone: (A) 0 and (B) 25 mM silver nitrate added. pH 7 HEPES buffer; 25 mM SDS; no organic modifier; 37 cm capillary, 30 cm effective length; 20 kV. Peaks: 1, thioxanthone and 2, dibenzothiophene.

by MEKC through the addition of 25 mM silver nitrate to the SDS-containing run buffer. Owing to their similar structures, these two vitamins represent a significant separation challenge in conventional liquid chromatography. The detection of these two compounds was performed at 280 nm, since the CAPS buffer absorbs at lower wavelengths, where the preferred detection of these compounds exists. Using this wavelength and because these compounds are very hydrophobic, thus minimizing their solubility in aqueous buffer, their detection is difficult. A good separation of the two compounds is obtained, however, upon the addition of 25 mM silver nitrate to the MEKC buffer, as shown in Figure 7B. The two compounds coeluted with the micelle under typical MEKC conditions (25 mM SDS), as shown in Figure 7A. Effect of Ag(I) on the Elution Window in MEKC. The effect of the addition of Ag(I) on the elution window in MEKC was also investigated. As noted previously, metal ion addition causes both a decrease in electroosmotic mobility and an increase in the electrophoretic mobility of the micelle (indicating a less negatively charged micellar surface due to the attraction of the positive metal ions).45,47 The presence of silver ions reduces the electrostatic repulsion between the polar head groups of the surfactant molecules, thus affecting the size and shape of the micelles formed and thereby altering the electrophoretic mobility of the micelle itself. In agreement with previous MEKC studies utilizing other metal ions, the addition of Ag(I) is shown to dramatically extend

Table 3. Data Showing the Effect of Silver Ion Addition on the Elution Window in MEKCa system

elution window

MEKC MEKC + 5 mM silver nitrate MEKC + 8 mM silver nitrate MEKC + 25 mM silver nitrate MEKC + 50 mM silver nitrate MEKC + 15% acetonitrile MEKC + 15% ACN + 5 mM silver nitrate MEKC + 15% MeOH MEKC + 15% MeOH + 5 mM silver nitrate MEKC + 15% MeOH + 25 mM silver nitrate MEKC + 15% MeOH + 50 mM silver nitrate MEKC + 15% n-propanol MEKC + 15% n-PrOH + 5 mM silver nitrate MEKC + 15% n-PrOH +25 mM silver nitrate MEKC + 15% n-PrOH +50 mM silver nitrate silver/sodium lauryl sulfate, 15 mM silver/sodium lauryl sulfate, 50 mM silver/sodium lauryl sulfate, 80 mM silver/sodium lauryl sulfate, 25 mM + 15% ACN silver/sodium lauryl sulfate, 25 mM + 15% MeOH silver/sodium lauryl sulfate, 25 mM + 15% n-PrOH silver/sodium lauryl sulfate, 80 mM + 15% ACN

2.4 2.6 2.6 3.3 4.8 3.4 4.4 4.4 5.8 >12 >12 4.3 4.7 >12 >12 2.6 3.1 3.4 6.8 5.9 6.0 6.8

a All data in this table were taken under the following conditions: MEKC utilizing 25 mM SDS, CAPS buffer (30 mM) at pH 10, 40 cm capillary (effective length), applied voltage of 15 kV.

Figure 7. MECC separation of vitamin D2 and vitamin D3: (A) 0 and (B) 25 mM silver nitrate added. pH 10 CAPS buffer; 25 mM SDS; no organic modifier; 47 cm capillary, 40 cm effective length; 15 kV. Peaks: 1, vitamin D2 and 2, vitamin D3.

the elution window. The reduction in electroosmotic flow with the addition of higher concentrations of metal ions is caused by a reduction in the ζ potential, defined as the charge created at the silica-solution interface under an applied field. Table 3 shows the values of the elution window or elution range, defined here as the ratio tmic/t0. As expected, the elution range is dramatically affected by the presence of Ag(I). A steady increase in the concentration of silver nitrate causes a gradual increase in the elution window. In going from 0% silver nitrate (simple MEKC conditions with no organic modifier) to 50 mM silver nitrate, the elution range doubles. It is well established that the addition of organic modifiers like methanol (MeOH), acetonitrile (ACN), or n-propanol (PrOH) can enhance the elution range through modification of the surface of the capillary, which results in changes in the ζ potential at the silica-solution interface (and thereby causes a reduction in electroosmotic flow velocity). Organic modifiers also tend to increase viscosity and reduce electroosmotic flow. Results here corroborate the previous observations. Additionally, however, we have shown that the combined addition of both organic modifiers and Ag(I) causes a further enhancement of the elution range, as demonstrated in Table 3. Furthermore, with a constant percentage of organic modifier (15%), the addition of higher concentrations of silver nitrate (25-50 mM) results in elution ranges greater than 12.

Finally, a mixed counterion surfactant of sodium/silver dodecyl sulfate was synthesized as described elsewhere.68 The percentage of silver, by weight, in the product was 29% ( 1%, as determined by a Volhard titration. Critical micelle concentrations of pure silver dodecyl sulfate have been reported to be 4.7 × 10-3 - 8.4 × 10-3 M, while that of pure SDS is 8.1 × 10-3 M.68 It is expected that the mixed counterion systems have cmc values slightly lower than that of pure SDS. Therefore, concentrations used in this study were assumed to be high enough to allow the formation of micelles. Changing the counterion on the dodecyl sulfate from Na+ to other metal ions can affect selectivity in MEKC separations and has been investigated by Foley and co-workers.69-71 The use of mixed Na+/Ag+ counterion dodecyl sulfate micelles in the present study did not show significantly different behavior than simply using silver nitrate as an additive to SDS-containing MEKC buffers, in terms of either elution window or retention characteristics. No attempts to optimize the synthesis for higher percentage of Ag+ were made. However, it is worth noting that the combination of the mixed Na+/Ag+ counterion dodecyl sulfate micelles and organic modifier results in a significant increase in the elution range, similar to the effect of the addition of Ag(I) to an SDS-containing buffer with organic modifier. In the case of ACN as the organic modifier in combination with 25 mM Na+/ Ag+ counterion dodecyl sulfate micelles, the elution range is twice that of 25 mM SDS with 15% ACN. CONCLUSION This study has shown that the use of Ag(I) as a buffer additive in both CZE and MEKC is an effective way to alter selectivity and retention mechanisms for separations of certain types of (68) Cline-Love, L. J.; Skrilec, M.; Habarta, J. G. Anal. Chem. 1980, 52, 754759. (69) Nielson, K. R.; Foley, J. P. J. Microcolumn Sep. 1993, 5, 347-360. (70) Nielson, K. R.; Foley, J. P. J. Microcolumn Sep. 1994, 6, 139-149. (71) Ahuja, E. S.; Foley, J. P. Anal. Chem. 1995, 67, 2315-2324.

Analytical Chemistry, Vol. 68, No. 3, February 1, 1996

423

solutes that can form weak complexes with Ag(I). This technique provides a complementary method to MEKC for the separation of these types of neutral solutes that can be difficult to separate by conventional MEKC. It should be noted that this study merely presents the utility of argentation electrophoresis as a viable alternative technique and provides examples of the range of applicability. No attempts were made to systematically optimize the various separations shown. In CZE, resolution of a pair of neutral sulfonamides is shown to improve greatly with an increase in Ag(I) concentration in the buffer. A CZE separation of nine sulfonamides was achieved by the addition of 50 mM silver nitrate to the buffer, thereby providing a form of secondary equilibria to alter the selectivity of the compounds that could not be separated under typical CZE conditions. The addition of Ag(I) was also shown to alter selectivity in MEKC separations through complexation phenomena in addition to the normal partitioning mechanisms generally governing MEKC

424

Analytical Chemistry, Vol. 68, No. 3, February 1, 1996

separations. The retention order of a mixture of five sulfonamides changed significantly when 25 mM Ag(I) was added to the SDScontaining buffer. Additionally, the use of Ag(I) as a buffer additive in MEKC allowed for the separation of a pair of S-containing heterocyclics and the separation of vitamins D2 and D3. Finally, as expected, the addition of Ag(I) to the buffer extended the elution range in MEKC. When combining both organic modifier, such as methanol, and Ag(I) as buffer additives in MEKC, the elution range more than doubles. ACKNOWLEDGMENT The authors are grateful for support of this research by NIH and AFOSR. Received for review August 8, 1995. Accepted November 10, 1995.X AC950802Q X

Abstract published in Advance ACS Abstracts, December 15, 1995.