CE and CEC. 1 Introduction. Electrophoresis 2004, 25, Xiaoming Chen Feng Qin Yueqi Liu Liang Kong Hanfa Zou

Transcription

1 Electrophoresis 2004, 25, Xiaoming Chen Feng Qin Yueqi Liu Liang Kong Hanfa Zou National Chromatographic Research & Analysis Center, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian, China Preparation of a positively charged cellulose derivative chiral stationary phase with copolymerization reaction for capillary electrochromatographic separation of enantiomers A positively charged chiral stationary phase (CSP) was prepared by chemically immobilizing cellulose 3,5-dimethylphenylcarbamate onto methacryloyldiethylenetriaminopropylated silica (MCDEAPS) via a radical copolymerization reaction. The prepared CSP was evaluated for enantiomer separation in nonaqueous capillary electrochromatography (CEC). Electroosmotic flow (EOF) generated on the prepared CSP could be significantly improved with introduction of positive charges into the CSP, and separation of enantiomers in CEC has been achieved with mobile phases of ethanol and hexane-ethanol, respectively. In addition, we investigated the solvent versatility of the immobilized CSP on enantioseparations in CEC and capillary liquid chromatography (CLC) due to the elimination of dissolution of chiral selector in a number of solvents. Chiral resolution of some enantiomers was improved by adopting tetrahydrofuran (THF) and chloroform as mobile-phase modifiers, respectively. Keywords: Capillary electrochromatography / Cellulose derivative / Chiral stationary phase / Copolymerization reaction / Enantiomeric resolution DOI /elps Introduction Correspondence: Dr. Hanfa Zou, National Chromatographic Research & Analysis Center, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian , China zouhfa@mail.dlptt.ln.cn Fax: Abbreviations: AIBN, a,a -azobisisobutyronitrile; CLC, capillary liquid chromatography; CSP, chiral stationary phase; DEA, diethylamine; DEAPS, diethylenetriaminopropylated silica; MCDEAPS, methacryloyldiethylenetriaminopropylated silica Separation of enantiomers in capillary electrochromatography (CEC) has attracted great attention due to its hybrid characteristics of high efficiency of capillary electrophoresis (CE) and high selectivity of high-performance liquid chromatography (HPLC) [1 5]. Cellulose derivative chiral stationary phases (CSPs) proved to be effective for resolution of a wide range of enantiomers in HPLC during the past two decades [6, 7]. In 1996, Francotte and Jung [8] reported enantiomer separation in open-tubular CEC by coating the capillary wall with cellulose 3,5-dimethylphenylcarbamate. Separation of enantiomers in packed CEC with cellulose derivative CSPs with aqueous mobile phases was reported [9 11]. Enantioseparations in nonaqueous CEC with cellulose derivative CSPs have been extensively studied by Chankvetadze and coworkers [12 20]. CEC enantioseparations of basic and bifunctional pharmaceuticals on Chiralcel OD-H and Chiralpak AD-H were recently reported [21]. Most of these studies focused on the evaluation of the coated-type CSPs prepared by physically coating cellulose derivatives onto silica gel. It is well-known that the application of coated-type CSPs is strictly limited because of the dissolution of cellulose derivatives in a number of solvents [22]. Therefore, preparation of chemically immobilized cellulose derivative CSPs for enantiomer separation in CEC might be interesting although the immobilized CSPs generally show lower chiral recognition ability and column capacity compared to coated CSPs. In our previous report, the chemically bonded types of cellulose phenylcarbamate CSPs were applied to CEC for enantiomer separation with both nonaqueous and aqueous mobile phases. The EOF generated could be improved by introduction of positive charges to the linkages [23, 24]. Recently, a method for preparation of bonded-type CSPs was developed by chemically immobilizing cellulose derivatives onto vinylized silica gel via a radical copolymerization reaction [25]. Immobilization could be efficiently attained in the presence of a small amount of a,a -azobisisobutyronitrile (AIBN) and the prepared CSPs generally showed relatively high column efficiency for HPLC resolution of enantiomers compared to other CSPs. Unfortunately, the newly prepared CSPs were almost neutral and the EOF generated might be poor in nonaqueous CEC. CE and CEC

2 2818 X. Chen et al. Electrophoresis 2004, 25, Thus, in this study we tried to adopt the same strategy previously reported to improve the EOF. Therefore, the newly developed radical copolymerization reaction could be extended for preparation of cellulose derivative CSPs applied to nonaqueous CEC. For this aim, the procedures of copolymerization reaction were modified to introduce a positively charged spacer into the CSP, which is expected to generate an anodic EOF in nonaqueous CEC. The chiral recognition ability of the prepared CSP was evaluated in CEC under nonaqueous mobile phases. The solvent effect on enantioseparations in CEC and capillary liquid chromatography (CLC) was also discussed. the procedures previously reported [24]. Then, 2.0 g of DEAPS was dispersed into a 50 ml flask containing 30 ml pyridine, and 300 ml methacryloyl chloride was slowly dropped within 1 min. The mixture was stirred and reacted at 707C for 5 h. Then, MCDEAPS was collected by centrifugation, washed with 300 ml methanol, and dried under vacuum. The results of elemental analyses of MCDEAPS are described as C%, 2.85; N%, 0.83 and H%, 0.38, respectively. 2 Materials and methods 2.1 Instrumentation and materials Separations throughout this study were performed on an Agilent capillary electrophoresis system (Hewlett-Packard, Waldbronn, Germany) with ChemStation software. A Waters 510 pump (Milford, MA, USA) was used to pack the capillary columns during our experiments. Elemental analyses were performed on the Elementar Vario EL (Hanau, Germany). Fused-silica capillaries (100 mm ID6365 mm OD) were obtained from Yongnian Optic Fiber Plant (Hebei, China), cellulose from Serva (Heidelberg, Germany), silica gel (Kromasil, 5 mm, 200 Å) from Akzo Nobel AB (Nacka, Sweden) and N -[3-(trimethoxylsilyl) propyl]diethylenetriamine from Aldrich Chemical Company (Milwaukee, WI, USA). Methacryloyl chloride was purchased from Acros (Geel, Belgium) and AIBN from Sanpu Chemical Company (Shanghai, China). Enantiomers of benzoin, warfarin, praziquantel, alprenolol, metoprolol, propranolol, and Tröger s base were all purchased from Sigma (St. Louis, MO, USA). Enantiomers of drug candidates A and B and ranolazine were from the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China). a-dimethyl dicarboxyl biphenyl derivatives of a-ddb1, a-ddb2, and a-ddb3 were obtained from the Lanzhou Institute of Chemical Physics, Chinese Academy of Sciences (Lanzhou, China). The molecular structures of the test enantiomers are presented in Fig Preparation of the positively charged CSP Preparation of methacryloyldiethylenetriaminopropylated silica (MCDEAPS) MCDEAPS was prepared as shown in Fig. 2. Diethylenetriaminopropylated silica (DEAPS) was prepared in advance with the reaction of bare silica gel and N -[3- (trimethoxylsilyl)propyl]diethylenetriamine according to Figure 1. Molecular structures of test enantiomers.

3 Electrophoresis 2004, 25, Enantioseparations by CEC Synthesis of cellulose 2,3-bis(3,5-dimethylphenylcarbamate)-6-methacrylate Cellulose 2,3-bis(3,5-dimethylphenylcarbamate)-6-methacrylate was prepared according to the procedures reported elsewhere [25]. Briefly, 0.5 g cellulose 2,3- bis(3,5-dimethylphenylcarbamate), prepared according to procedures as reported in [26], was dissolved in 30 ml pyridine, and 200 ml methacryloyl chloride was slowly added into the solution within 1 min. The mixture was then allowed to react at about C for 12 h in the presence of N 2. The solution was poured into a 500 ml beaker containing 300 ml methanol. Finally, cellulose 2,3-bis(3,5-dimethylphenylcarbamate)-6-methacrylate was collected by centrifugation and completely washed with 300 ml methanol Radical copolymerization reaction Cellulose 2,3-bis(3,5-dimethylphenylcarbamate)-6-methacrylate (0.5 g) was dissolved in THF (15 ml), then 2.0 g of MCDEAPS was added. After stirring for 20 min, the solvent of THF was evaporated, thus the synthesized cellulose derivative was coated onto the modified silica gel. The silica gel coated with cellulose derivative was then placed into a 50 ml flask containing 50 mg AIBN. The mixture was stirred and reacted at about 1007C for 2 h. After that, 30 ml pyridine was poured into the flask to completely dissolve the residues. The product was obtained by centrifugation and washed with 30 ml THF and 300 ml methanol, respectively. Thus, the CSP containing secondary amino groups onto the linkage spacer was obtained (Fig. 2). The prepared CSP was characterized by elemental analyses as C%, 10.3; N%, 1.43 and H%, 0.64, respectively. The noncharged CSP, which has been previously prepared for HPLC resolution of enantiomers, was also obtained with the above procedures except that MCDEAPS was displaced by vinylized silica gel prepared by the reaction of bare silica gel and g-(trimethoxysilyl)propyl methacrylate [25]. The difference between the positively charged CSP and the noncharged CSP is only that the former contains the secondary amino groups in the spacer between cellulose derivative and silica gel. 2.3 Preparation of packed capillary columns CEC columns were prepared by a slurry packing method previously reported [23]. Briefly, the slurry was prepared by dispersing the CSP in ethanol followed by sonication for 20 min. The slurry packing of the capillary column was carried out on a Waters 510 pump. After packing, the frits were made by sintering the CSP on a small portion of the packed capillary, and the residue particles behind the outlet frit were then flushed out. Finally, a detection window was created just behind the outlet frit. 2.4 Separation conditions Enantiomer separationswere carried out under the following conditions (unless otherwise stated): All enantioseparations were performed with short-end column, in Figure 2. Radical polymerization reactions for preparation of the positively charged CSP.

4 2820 X. Chen et al. Electrophoresis 2004, 25, which a fused-silica capillary of 32.0 cm with a packed length of 8.0 cm was used for CEC and CLC experiments. The temperature was kept at 257C for all experiments and the detection wavelength was set at 214 nm. The injections were made with a pressure of 4 bar to the inlet vial for 0.10 min for CEC and CLC experiments. During our study, a negative voltage indicated that the cathode was placed at the inlet side and the anode at the outlet side of the column, and ethanol was selected as t 0 marker. The buffer system consisted of 14.2 mm acetic acid and 1.4 mm diethylamine (DEA) under various mobile-phase conditions. A pressure was only adopted to the inlet side for the experiments in CLC and pressure-assisted CEC. 3 Results and discussion 3.1 Characterization of EOF EOF (m eo ) generated on the prepared CSP is important for successful enantiomer separation in CEC. Thus, we first evaluated the characteristics of m eo on the positively charged CSP with mobile phase of ethanol containing 14.2 mm acetic acid and 1.4 mm DEA with an apparent ph of The positively charged CSP presented an anodic EOF from cathode to anode when a negative voltage was applied. It can be explained that the secondary amino groups on the CSP could form positive charges under a low ph, thus generating the anodic EOF. The magnitude of m eo was 4.55 cm 2?kV 21?min 21 when a voltage of 215 kv was applied, which was very close to that reported previously on a similarly prepared CSP under the same separation conditions (about 4.80 cm 2?kV 21?min 21 ) [24]. EOF on the noncharged CSP was also measured under the same conditions. However, it was found that the t 0 marker was difficult to be eluted within 60 min under both negative or positive voltage. This result might suggest that the residual silanol groups on the noncharged CSP were not sufficient to generate a considerably high EOF to drive the liquid flow under our experimental conditions. Therefore, the introduction of charged groups into the CSP is necessary to improve the EOF, which is a prerequisite for successful enantioseparations in CEC. 3.2 Enantiomer separation in CEC with classic mobile phases Cellulose derivative-based CSPs have been widely used for enantiomer separation in HPLC, and most of the separations have been carried out with nonaqueous mobile phases, such as a hexane-alcohol system [6, 7]. In our previous reports, the chemically bonded types of cellulose phenylcarbamate CSPs have been applied for CEC separations, an hexane-alcohol system as mobile phase indicating the potential for seeking wider application in CEC with these kinds of CSPs [23, 24]. In this study, since the noncharged CSP generated a poor EOF in nonaqueous mobile phases, it is hardly used for CEC separations. Therefore, enantioseparations throughout this study were carried out on the positively charged CSP under low-ph mobile phases of ethanol and hexane-ethanol system. Electrochromatographic data for enantiomer separation in CEC with a mobile phase of pure ethanol are shown in Table 1. The R s value was calculated as 1.18(t 2 2t 1 )/(w 1/2(1) 1 w 1/2(2) ), respectively, where t 1 and t 2 refer to the retention times for the first and second eluted enantiomers, and w 1/2(1) and w 1/2(2) are the peak widths at half height for the first and second eluted enantiomers, respectively. As can be seen, five enantiomers were successfully separated with a negatively applied voltage. Among them, Tröger s base, ranolazine, and drug candidate A were completely resolved, while praziquantel and drug candidate B were partially separated. It should be pointed out that the basic enantiomers of ranolazine and drug candidate B were only eluted with the mode of pressureassisted CEC, possibly due to the opposition of electrophoretic mobility to the anodic EOF. A typical electrochromatogram for the separation of Tröger s base is shown in Fig. 3. The enantiomers were eluted within 7 min with an applied voltage of 215 kv, which also suggested that the EOF generated on the positively charged CSP was considerable higher in this case. In addition, better resolution for this enantiomer was observed in terms of R s compared to our previous report on a similar CSP [24]. Table 2 lists the electrochromatographic data for enantiomer separation in CEC with a hexane-ethanol system as mobile phase. In this case, more enantiomers have been resolved with different chromatographic conditions as listed in Table 2. However, it was observed that the EOF was somewhat decreased by addition of the nonpolar solvent hexane into the mobile phase compared Table 1. Electrochromatographic separation of enantiomers on the positively charged CSP with mobile phase of ethanol Racemate Applied voltage (kv) t 1 (min) t 2 (min) Tröger s base Praziquantel Ranolazine bar Drug candidate A Drug candidate B bar a R s

5 Electrophoresis 2004, 25, Enantioseparations by CEC 2821 separated within 15 min, respectively, whereas the other enantiomers were eluted within acceptable elution times. Figure 3. Separation of Tröger s base with mobile phase of ethanol. Electrochromatographic conditions as in Table 1. Plate numbers were calculated per meter. to that with pure ethanol as mobile phase. For example, the magnitude of m eo was 1.34 cm 2?kV 21?min 21 with the mobile phase of hexane-ethanol (50/50) by applying a voltage of 215 kv. Thus, a higher voltage should be applied in this case in order to shorten the analysis time. As shown in Table 2, most of the test enantiomers were separated with high a and R s values under the applied voltages. A little higher R s was found compared to that obtained on a similar CSP previously prepared, such as Tröger s base, praziquantel, and drug candidate A [24]. Typical electrochromatograms for the resolution of Tröger s base and drug candidate A are presented in Fig. 4. Both enantiomers were completely During our experiments, we tried to resolve some basic enantiomers with a mobile phase of hexane-ethanol, such as ranolazine, alprenolol, metoprolol, and propranolol. Again, it was found that they were difficult to be eluted on the positively charged CSP both under a positive or negative voltage. The reason might be also due to the opposition of electrophoretic mobility to the anodic EOF generated under the experimental conditions. Thus, pressure-assisted CEC was adopted for resolution of these basic enantiomers (Table 2). It can be seen that these basic enantiomers could be separated with various pressures into the inlet side of the capillary. Typical chromatograms for separation of ranolazine and propranolol are shown in Fig. 5. The enantiomers could be resolved by pressure-assisted CEC within 5 min by applying pressures of 12 and 10 bar, respectively. The other basic enantiomers could be eluted within several minutes as well. Although the test enantiomers could be resolved on the prepared CSP in nonaqueous CEC, the average value of the plate numbers for the eluted peaks is only about plates/m, which is somewhat lower and similar to that previously reported on another positively charged CSP prepared with a bifunctional reagent of tolylene-2,4- diisocyanate [24]. One of the possible reasons is attributed to the chromatographic characters of the positively charged CSP, on which the column efficiency was decreased after the introduction of the positively charged Table 2. Electrochromatographic separation of enantiomers on the positively charged CSP with mobile phase of hexane-ethanol system Racemate Applied voltage (kv) Mobile phase t 1 (min) t 2 (min) a R s Tröger s base 225 a Benzoin 225 a Praziquantel 225 b Warfarin 220 b a-ddb1 225 b a-ddb3 225 b Drug candidate A 230 c Drug candidate B bar c Ranolazine bar c Alprenolol bar d Metoprolol bar c Propranolol bar c Mobile phases: a, hexane-ethanol (60/40); b, hexane-ethanol (55/45); c, hexane-ethanol (50/50); d, hexane-ethanol 65/35

6 2822 X. Chen et al. Electrophoresis 2004, 25, Figure 4. Separation of (a) Tröger s base and (b) drug candidate A with hexane-ethanol as mobile phase. Electrochromatographic conditions as in Table 2. Figure 5. Separation of (a) ranolazine (a) and (b) propranolol with pressure-assisted CEC. Electrochromatographic conditions as in Table 2. spacer. The narrow pore diameter of the silica used for preparation of the CSP may also result in a low performance. In addition, the technique for frit preparation might be another reason responsible for the low column performance in CEC. The successful CEC enantioseparations achieved on the prepared CSP have demonstrated that the EOF could be effectively enhanced under nonaqueous mobile phases with the introduction of positive charges into the CSP prepared with the radical copolymerization reaction. Thus, it might be an alternative strategy to modify some traditional CSPs, which have been widely used in HPLC but were not adaptable in CEC, in order to make them available in CEC for separation of enantiomers. 3.3 Solvent effect with the mobile-phase modifiers THF and chloroform As we know, the coated types of cellulose derivativebased CSPs, which are prepared by physically coating cellulose derivatives onto silica gel, are limited to be used in practice because of the solubility of cellulose derivatives in a number of solvents, such as THF and chloroform [22]. Thus, chromatographic separation of some enantiomers might be problematic due to the rigid restriction of solvents in the mobile phases. However, the dissolution of cellulose derivatives is greatly eliminated on the chemically immobilized CSPs, therefore, those solvents, generally not compatible to the coated types of CSPs, can be used as mobile-phase modifiers on the chemically immobilized CSPs. These issues have been discussed in

7 Electrophoresis 2004, 25, Enantioseparations by CEC 2823 HPLC resolution of enantiomers [22, 26]. It was pointed out that some enantiomers, poorly recognized on the coated types of CSPs, could be well resolved on the chemically immobilized CSPs with addition of THF and chloroform into the mobile phases, respectively. Thus, we tried to investigate the solvent effect on the prepared CSP in nonaqueous CEC. A typical example is shown in Fig. 6 for separation of praziquantel with THF as a mobile-phase modifier. It can be seen that the EOF was slightly changed judged from the t 0 marker with addition of 5% v/v THF into the mobile phase. However, the elution time for the eluted enantiomers was clearly shortened. Although the enantioselectivity was insignificantly changed, the magnitude of R s increased from 2.19 to 2.67, and N 1 and N 2 were also improved to various degrees, respectively, with the mobile-phase modifier THF. Better resolution in CEC was obviously obtained after addition of a small portion of THF to the mobile phase. hexane-ethanol (70/30), while the a- and R s values were greatly improved with the mobile phase hexane-ethanolchloroform (85/15/5). Unfortunately, the EOF generated on the prepared CSP was much more poor with the mobile phases hexane-ethanol-thf (80/15/5) and hexane-ethanol-chloroform (85/15/5) due to the increase of the nonpolar solvent hexane in the mobile phases [23]. We tried to increase the volume of ethanol in the mobile phases in order to generate a higher EOF but both of the enantiomers could not be resolved under these conditions. Thus, electrochromatograms for separation of both enantiomers with the mobile-phase modifiers THF During our previous HPLC experiments it was found that chiral separation of racemates of a-ddb2 and drug candidate B was also markedly affected by the solvents of THF and chloroform, respectively. Thus, both enantiomers were resolved in the mode of CLC, and the obtained chromatograms are presented in Fig. 7. The enantiomer of a-ddb2 almost could not be recognized under the classic mobile phase hexane-ethanol (80/20), while it was completely resolved with addition of 5% THF into the mobile phase. The enantiomer of drug candidate B was only partially separated in the classical mobile phase Figure 6. Separation of praziquantel with addition of THF to mobile phase. Mobile phases: (I) hexane-ethanol (60/40); (II) hexane-ethanol-thf (60/40/5). Figure 7. Enantiomer separation in CLC with addition of THF and chloroform to the mobile phases. (a) Resolution of a-ddb2 by applying a pressure of 11 bar with mobile phases of (I) hexane-ethanol (80/20) and (II) hexane-ethanol-thf (80/15/5). (b) Resolution of drug candidate B by applying a pressure of 12 bar with mobile phases of (I) hexane-ethanol (70/30) and (II) hexane-ethanol-chloroform (85/15/5).

8 2824 X. Chen et al. Electrophoresis 2004, 25, and chloroform could not be generated in this study. We have tried to find more such examples but failed due to the limited enantiomers available in our lab. Anymore, the potential for enantiomer separation in CEC on the bonded type of CSPs by using other solvents in the mobile phase is indicated. 4 Concluding remarks A radical polymerization reaction was applied to prepare the chemically bonded type of cellulose phenylcarbamate CSP with a positively charged spacer. The positively charged CSP generated a considerably higher EOF from cathode to anode in nonaqueous CEC. Therefore, separation of enantiomers in CEC on the prepared CSP could be performed with mobile phases consisting of ethanol and hexane-ethanol. Most of the neutral enantiomers could be resolved within 15 min, while the basic enantiomers could be resolved in pressure-assisted CEC within 5 min. Resolution for enantiomers of praziquantel, a-ddb2, and drug candidate B have been markedly improved in CEC and CLC with the mobile-phase modifiers of THF and chloroform, respectively, which indicated the possibility in the search for separation of enantiomers on the chemically immobilized CSP with different solvents as modifiers of mobile phases. The financial support from the Natural Science Foundation of China (No ) and the Knowledge Innovation Program of Dalian Institute of Chemical Physics to Dr. Hanfa Zou is gratefully acknowledged. Received February 13, References [1] Wistuba, D., Schurig, V., J. Chromatogr. A 2000, 875, [2] Lämmerhofer, M., Svec, F., Fréchet, J. M. J., Lindner, W., Trends Anal. Chem. 2000, 19, [3] Chankvetadze, B., J. Sep. Sci. 2001, 24, [4] Kang, J., Wistuba, D., Schurig, V., Electrophoresis 2002, 23, [5] Scriba, G., Electrophoresis 2003, 24, [6] Okamoto, Y., Yashima, E., Angew. Chem. Int. Ed. 1998, 37, [7] Yashima, E., J. Chromatogr. A 2001, 906, [8] Francotte, E., Jung, M., Chromatographia 1996, 42, [9] Krause, K., Girod, M., Chankvetadze, B., Blaschke, G., J. Chromatogr. A 1999, 837, [10] Otsuka, K., Mikami, C., Terabe, S., J. Chromatogr. A 2000, 887, [11] Mayer, S., Briand, X., Francotte, E., J. Chromatogr. A 2000, 875, [12] Meyring, M., Chankvetadze, B., Blaschke, G., J. Chromatogr. A 2000, 876, [13] Girod, M., Chankvetadze, B., Blaschke, G., J. Chromatogr. A 2000, 887, [14] Chankvetadze, B., Kartozia, I., Breitkreutz, J., Okamotom, Y., Blaschke, G., Electrophoresis 2001, 22, [15] Fanali, S., Catarcini, P., Blaschke, G., Chankvetadze, B., Electrophoresis 2001, 22, [16] Girod, M., Chankvetadze, B., Blaschke, G., Electrophoresis 2001, 22, [17] Chankvetadze, B., Kartozia, I., Breitkreutz, J., Girod, M., Knobloch, M., Okamoto, Y., Blaschke, G., J. Sep. Sci. 2001, 24, [18] Chankvetadze, B., Kartozia, I., Breitkreutz, J., Okamoto, Y., Blaschke, G., Electrophoresis 2001, 22, [19] Chankvetadze, B., Kartozia, I., Okamoto, Y., Blaschke, G., J. Sep. Sci. 2001, 24, [20] Chankvetadze, L., Kartozia, I., Yamamoto, C., Chankvetadze, B., Blaschke, G., Okamoto, Y., Electrophoresis 2002, 23, [21] Mangelings, D., Hardies, N., Maftouh, M., Suteu, C., Van der Heyden, Y., Electrophoresis 2003, 24, [22] Franco, P., Senso, A., Oliveros, L., Minguillón, C., J. Chromatogr. A 2001, 906, [23] Chen, X., Zou. H., Ye. M., Zhang. Z., Electrophoresis 2002, 23, [24] Chen, X., Jin, W., Qin, F., Liu, Y., Zou. H., Guo, B., Electrophoresis 2003, 24, [25] Chen, X., Qin, F., Liu, Y., Huang, X., Zou. H., J. Chromatogr. A 2004, in press. [26] Yashima, E., Fukaya, H., Okamoto, Y., J. Chromatogr. A 1994, 677,