Journal of Chromatography A, 1166 (2007)

Transcription

1 Journal of Chromatography A, 1166 (2007) Comparison of two sample preconcentration strategies for the sensitivity enhancement of flavonoids found in Chinese herbal medicine in micellar electrokinetic chromatography with UV detection Jinhua Zhu a,b, Kai Yu a,b, Xingguo Chen a,b,, Zhide Hu a,b a State Key Laboratory of Applied Organic Chemistry, Lanzhou University, Lanzhou , China b Department of Chemistry, Lanzhou University, Lanzhou , China Received 19 April 2007; received in revised form 2 August 2007; accepted 9 August 2007 Available online 12 August 2007 Abstract Two on-column preconcentration techniques named stacking with reverse migrating micelles (SRMM) and anion selective electrokinetic injection and a water plug-sweeping with reverse migrating micelles (ASIW sweep RMM) were used and compared for concentration and separation of flavonoids in Chinese herbs using reverse migration micellar electrokinetic chromatography (RM-MEKC). The optimal background electrolyte (BGE) used for separation and preconcentration was a solution composed of 20 mm phosphoric acid (H 3 PO 4 ) 100 mm sodium dodecyl sulfate (SDS) 20% (v/v) acetonitrile (ACN) buffer (ph 2.0), the applied voltage was 15 kv. To achieve reasonable results of the two techniques, the conditions which affected preconcentration were examined. A comparison of used techniques with normal hydrodynamic injection (5 s), concerning enhancement factors and limits of detection (LODs) was presented. Under the optimum stacking conditions, about and fold improvement in the detection sensitivity was obtained for SRMM and ASIW sweep RMM, respectively, compared to usual hydrodynamic sample injection (5 s). The LODs (S/N = 3) for SRMM and ASIW sweep RMM in terms of peak height, can reach down to g/ml for hesperetin and g/ml for nobiletin, respectively. Finally, the amounts of the six flavonoids in extract of Fructus aurantii Immaturus were successfully determined using ASIW sweep RMM. The six analytes were baseline separated with sample matrix under the optimum ASIW sweep RMM conditions and the experimental results showed that preconcentration was well achieved after the dilution of sample solutions Elsevier B.V. All rights reserved. Keywords: Sample preconcentration; Capillary electrophoresis; Reverse migrating micelles; Micellar electrokinetic chromatography; Flavonoids 1. Introduction Capillary electrophoresis (CE) has matured over the past few years into a powerful analytical tool used in diverse fields, and is considered orthogonal or complementary to high-performance liquid chromatography (HPLC). Since micellar electrokinetic chromatography (MEKC) was developed and introduced into the research field of CE by Terabe et al. [1], the application area of CE has been remarkably broadened as both ionic and neutral analytes can be separated by MEKC [2]. However, one serious drawback of CE with UV detection is the low concentration sensitivity which is associated with the lim- Corresponding author at: Department of Chemistry, Lanzhou University, Lanzhou , China. Tel.: ; fax: address: chenxg@lzu.edu.cn (X. Chen). ited optical path length for on-line UV detection and the small amount of sample that can be injected into the capillary. Several strategies have been proposed to improve detection limits. It is possible to use more sensitive detection methods (such as laser-induced fluorescence (LIF), electrochemical, MS, fluorescence, and chemiluminescence detectors), or use Z-shape and multi-reflection cell to increase optical path length [3,4]. However, instruments equipped with these detectors are expensive or require modifications of CE system. In response to this, several preconcentration approaches have been developed for the sensitivity enhancement. One is an off-line preconcentration method based on solid- or liquidphase extraction prior to CE analysis [5 7]. However, these procedures are tedious. Recently, there has been considerable interest in developing on-line preconcentration techniques in CE [8 23]. They offer several advantages over off-line preconcentration methods, such as simplicity, high enrichment factors, and /$ see front matter 2007 Elsevier B.V. All rights reserved. doi: /j.chroma

2 192 J. Zhu et al. / J. Chromatogr. A 1166 (2007) automation. Furthermore, separation and preconcentration can be performed in a single capillary without modification of existing instruments [24]. In general, these methods are designed to compress long analyte bands within the capillary thereby increasing the sample volume that can be injected, without the loss of CE efficiency. Preconcentration of charged as well as neutral analytes is possible using MEKC [25,26]. Two on-line sample concentration techniques, sample stacking and sweeping, are known to be attractive approaches for enhancement of the concentration sensitivity in MEKC [8,15,22,26]. Stacking is based on the electric field strength differences between the sample zone and the background electrolyte (BGE). In brief, sample stacking occurs at the boundary between the high electric field sample zone and low electric field BGE zone. Samples are prepared in low-conductivity matrix [26]. It results from a rapid change in migration velocity of micelles when passing from one region to another. Quirino and Terabe verified and explained the effect of experimental parameters on stacking model [12]. The stacking efficiency had a strong relationship with the retention factors of analytes and the velocity of micelles. High retention factor compounds travel faster than low retention factor compounds. If the micelles migrated too quickly, for analytes which had low retention factors, it would lead to the sample zone partly present in the micelles to be stacked and partly retain in buffer, which would result in broad or splitted peaks [12]. If the micelles migrated too slowly, for analytes that had high retention factors, the sample zone that already stacked by micelles would be retarded. Due to dispersive effect, broad peaks would appear, too. Therefore, the selection of appropriate BGE conditions and the optimization of preconcentration parameters were quite important. When stacking is exercised in MEKC, reverse migration micellar electrokinetic chromatography (RM-MEKC) which is characterised by micelles moving faster than the electroosmotic flow (EOF) is always used [22,12]. Some papers showed that sample stacking techniques in RM-MEKC mode using low ph phosphate buffers were very good [12,14]. In RM-MEKC, when the stacking process is performed using hydrodynamic injection, it is named stacking with reverse migrating micelles (SRMM). SRMM is performed by hydrodynamical injection of samples which are prepared in low-conductivity matrix into the capillary for long periods of time. The sweeping phenomenon is also partly responsible for the focusing effect in SRMM [27]. To effect stacking and separation, voltages are simply applied at negative polarity after injection, with the separation solution in the anodic and cathodic vials. Since the negative polarity is applied at the inlet, the sample matrix was slowly pushed out of the capillary by the weak EOF. Only negative polarity is needed to focus and separate the analytes without the polarity-switching step in SRMM. Another on-line preconcentration technique, sweeping, was originally demonstrated by Terabe et al. [15]. Sweeping is defined as a phenomenon where analytes are picked up and concentrated by the pseudostationary (PS) phases that penetrates the sample zone devoid of PS. It is independent of the BGE and sample zone conductivities and effective for both charged and uncharged solutes. The only requirement for sweeping to take place is the absence of the PS in the sample matrix. Sweeping can be exercised in RM-MEKC. When low ph phosphate buffers were used in RM-MEKC, the approach was named sweeping with reverse migrating micelles (sweep-rmm). Analytes can be injected into the capillary hydrodynamically or electrokinetically. If electrokinetical injection mode is used, analytes can be selectively injected under proper polarity. Anion selective electrokinetic injection-sweeping (ASI-sweep) is a combination of two on-line preconcentration techniques, sample stacking with electrokinetic injection and sweeping. In order to effectively perform sample stacking with electrokinetic injection and succedent sweeping, the sample has to be prepared in a low-conductivity matrix that avoid of PS [26]. Generally, a short plug of water is introduced hydrodynamically before the electrokinetic sample injection. The water plug provides a high electric field at the tip of the capillary in sample stacking by electrokinetic injection, which will eventually improve the sample stacking procedure [26]. When this whole procedure is carried out in RM-MEKC, it could be called anion selective electrokinetic injection and a water plug-sweeping with reverse migrating micelles (ASIW sweep RMM). With this method, first, a short plug of water was hydrodynamically injected after the capillary had been conditioned with acidic micellar BGE. In the next step, the anion sample prepared in a low-conductivity solution is injected electrokinetically at negative polarity. The sample anions enter the capillary through the water plug with high velocities. Once the sample ions reach the interface between the water and BGE, they will slow down and focus at this interface. Finally, the separation voltage is applied at negative polarity with the micellar BGE in the inlet vial. Micelles enter the capillary and sweep the stacked analytes. At last, the stacked anions are completely swept by the micelles and are separated by MEKC in the reverse migration mode. In the whole procedure, only negative polarity is needed and there is no polarity switching, either. Fructus aurantii Immaturus (Zhishi in Chinese) is the dried unripen fruit of Citrus aurantium L. It contains abundant flavonoids, such as hesperidin, nobiletin, tangeretin, hesperetin, naringenin, naringin, narirutin, neohesperidin, etc. These flavonoids have been discovered for antioxidative [28,29], anticarcinogenic [30 32], anti-inflammatory [33,34] effects, etc. Flavonoids are usually used as chemical markers for the quality control of natural plants and fruits because of their ubiquity and multiplicity in fruits, and specificity to different varieties and even cultivars. So it is necessary to establish a suitable analytical method to evaluate these flavonoids for pharmaceutical and herbal preparations. In the past, methods based on liquid chromatography (LC) [35], HPLC [36 38] and CE [39 42] have been reported for this purpose. However, since it is easy to obtain their standards and their contents are relatively high in samples, hesperidin, hesperetin, naringenin, and naringin were usually determined and used as pharmaceutical quality control in the past. And those methods mentioned above also had only been used to analyze part of the flavonoids, generally three or four and the detection limits were not considered. Because nobiletin and tangeretin were also flavonoids contained in F. aurantii Immaturus, and there were so many promising properties of them [43], so it is better to analyze them together with other flavonoids in F. aurantii Immaturus. To our best knowledge,

3 J. Zhu et al. / J. Chromatogr. A 1166 (2007) there are no reports published on the simultaneous determination of hesperidin, nobiletin, tangeretin, hesperetin, naringenin, and naringin in F. aurantii Immaturus. So, the six flavonoids commonly present in F. aurantii Immaturus were chosen in this study as model analytes. Firstly, investigated methods are easy to perform, and they do not require specialized personnel and skills and could simplify the entire procedure of determination. Secondly, standard CE system equipped with UV detector is relatively cheap alternative comparing to other detectors like LIF and MS. And the two techniques are achievable with standard CE apparatus with UV detection. And also, there is no polarity switching involved in the two methods, which is not always possible in most commercial CE instruments. Impressive concentration capabilities were confirmed by already published results. Thirdly, because of so many promising properties of flavonoids in F. aurantii Immaturus and the insufficiency of available data concerning preconcentration of flavonoids compounds in F. aurantii Immaturus by stacking or sweeping, therefore, pursue for simplest and most economic, yet reliable methods for the quantitative and qualitative analysis of flavonoid markers in F. aurantii Immaturus is desirable. So the major aim of this work is to select six flavonoids (their structures were shown in Fig. 1) in F. aurantii Immaturus as model analytes to evaluate comparatively the performance of two on-line preconcentration techniques, SRMM and ASIW sweep RMM, respectively, to increase the concentration sensitivity in CE with UV detection. Optimum conditions for SRMM and ASIW sweep RMM were obtained. Then, the enhancement factors and parameters of the two techniques were systematically evaluated and compared. Finally, the extract of F. aurantii Immaturus was analyzed and determined by ASIW sweep RMM stacking technique which provided better sensitivity enhancement. Fig. 1. Chemical structures of the analytes.

4 194 J. Zhu et al. / J. Chromatogr. A 1166 (2007) Experimental 2.1. Apparatus All CE experiments were performed using a Quanta 4000 instrument (Waters, Milford, MA, USA). Separation and stacking were carried out on a fused-silica capillary (Yongnian Photoconductive Fiber Factory, Hebei, China) of 60.0 cm (effective length 52.5 cm) 75 m i.d. The UV detector was set at 214 nm. Data acquisition was carried out with a Maxima 820 chromatography workstation. A PHS-3B acidity meter (Shanghai Precision & Scientific Instrument Co., Shanghai, China) was used for the ph measurement. The ph was adjusted with 1 M NaOH or HCl Chemicals and materials Standards of flavonoids (tangeretin (1), nobiletin (2), hesperetin (3), naringenin (4)) were purchased from Shanxi Huike Botanical Development Co. (Xi an, China). Hesperidin (5) and naringin (6) were obtained from the National Institute for the Control of Pharmaceutical and Biological Products (Beijing, China). F. aurantii Immaturus were purchased from Zhongyou drugstore (Lanzhou, China). Sodium dodecyl sulfate (SDS) was purchased from Shanghai Chemical Reagent Company (Shanghai, China). Phosphoric acid (H 3 PO 4 ), methanol (MeOH), acetonitrile (ACN), sodium hydroxide and hydrochloric acid were purchased from Tianjin Chemical Reagent Factory (Tianjin, China). All chemicals and solvents were of analytical reagent grade and were used without further purification. Distilled water was used throughout Solutions and sample preparation Stock standard solutions were prepared in MeOH to give concentration of 1240 g/ml for 1, 950 g/ml for 2, 1000 g/ml for 3, 5, and 6, and 1140 g/ml for 4. A mixed standard solution of the six flavonoids was prepared weekly by mixing the six stock solutions and distilled water with appropriate volumes. The working standard solutions were prepared by diluting the stock standard solutions with distilled water. All the standard solutions were stored at room temperature when they were prepared. The run buffer solution was prepared daily by mixing appropriate volumes of 0.2 M H 3 PO 4, 0.2 M SDS, ACN and distilled water. All the solutions for CE were filtered through a 0.45 m cellulose acetate membrane filter (Shanghai Xinya Purification Apparatus Factory, Shanghai, China) prior to use. A 1.0 g sample of pulverized F. aurantii Immaturus was immersed in 10 ml MeOH at room temperature for 12 h, after that, it was extracted in an ultrasonic bath for 60 min and followed by centrifugation at 1500 g for 5 min. Then the solution was adopted as the sample stock solution. The sample stock solution was diluted 800-fold for the determination of 5 and 6 in the sample and 20-fold for the determination of other four flavonoids in the sample Operation procedure Prior to use, the capillary was pretreated with the following cycles: distilled water for 10 min, 0.1 M NaOH for 10 min and distilled water for 10 min. At the beginning of each day, the capillary was equilibrated by rinsing with 0.1 M NaOH for 5 min, distilled water for 5 min, and then the BGE for 5 min. The capillary was pre-conditioned between consecutive analyses with distilled water for 2 min, followed by methanol for 2 min, 0.1 M NaOH for 3 min, distilled water for 2 min and finally the BGE for 3 min. When ASIW sweep RMM stacking technique was used, the procedure was carried out according to the previous report described by Wang et al. [44]. When SRMM was performed, the procedure described by Quirino and Terabe [12] was adopted. Other experimental conditions were specified in the figures or text Stacking enhancement factors calculation Stacking enhancement factors in terms of peak height, peak area and limits of detection (LODs), SE height,se area, and SE LOD, respectively, were calculated using these equations: SE height = H stack H, SE area = A stack A, SE LOD = LODs 5, LODs stack where H stack, A stack, and LODs stack are the peak height, peak area and LODs obtained when preconcentration was performed, H, A, and LODs 5 are the peak height, peak area and LODs obtained from normal hydrodynamic sample injection (5 s). LODs stack and LODs 5 were calculated from respective calibration curves. The values of LODs 5 and LODs stack were obtained through the slopes of respective calibration curves divided by three times of the standard deviation of baseline. Alternatively, SE LOD was calculated as a ratio of the LOD values obtained by the normal hydrodynamic injection and the ones for the preconcentration techniques. 3. Results and discussion 3.1. Optimization of the separation conditions A conventional hydrodynamic sample injection (5 s, 10 cm) and an 15 kv applied voltage were used to optimize the micellar BGE solution composition. The separation was optimized by studying the effects of the electrolyte ph and the concentrations of SDS and organic modifiers on the resolution between adjacent peaks. The optimum BGE for separation of six flavonoids was as follows: 20 mm H 3 PO 4 buffer (ph 2.00) containing 100 mm SDS and 20% (v/v) ACN. A typical electropherogram was shown in Fig. 2A. The migration sequence of analytes coincided with that of their retention factors. The higher the retention factors, the longer they retained in the micelles which had faster velocity than that of EOF. So, the shorter were the migration times.

5 J. Zhu et al. / J. Chromatogr. A 1166 (2007) Fig. 3. Effect of injection time on SRMM. The sample was hydrodynamically (10 cm) injected for s. Buffer and other conditions are the same as those in Fig. 2B. Concentrations of analytes were in the same as in Fig. 2B. few literatures that reported the comparison of different on-line concentration strategies to analyze neutral or ionic compounds in MEKC [24,45 49], but the combination and comparison of the two techniques to stack flavonoids was applied for the first time in this paper using an acidic micellar BGE solution. The theory of the two stacking techniques had been introduced in Section 1. To obtain the optimum preconcentration results, parameters that affected enhancement factors needed to be optimized. Fig. 2. Electropherograms of standard solutions of the six flavonoids using different procedures: (A) normal hydrodynamic sample injection (5 s, 10 cm). (B) SRMM: hydrodynamic sample injection at 10 cm for 100 s. (C) ASIW sweep RMM: water plug was hydrodynamically injected into the capillary at 10 cm for 2 s and then, the sample was loaded electrokinetically for 15 s under a voltage of 10 kv. BGE, 20 mm H 3 PO 4 buffer (ph 2.0) containing 100 mm SDS and 20% (v/v) ACN; applied voltage: 15 kv. The concentration of the analytes were 6.2 g/ml for 1, 9.5 g/ml for 2, 6.5 g/ml for 3, 11.4 g/ml for 4, 15 g/ml for 5, 6 in (B); 6.2 g/ml for 1, 4.75 g/ml for 2, 10 g/ml for 3, 5, 6, 11.4 g/ml for 4 in (A) and (C). For other experimental conditions, see text Optimization of stacking conditions SRMM and sweeping are two on-column preconcentration techniques in MEKC that have previously been reviewed by Kim and Terabe [26]. To our best knowledge, there were only a SRMM In order to achieve maximum concentration factors, different hydrodynamic sample injection times between 30 and 150 s (10 cm) were tested. With the injection time increasing from 30 to 100 s, the peak heights increased as expected. However, further increase of injection time to 130 s, peak heights of the test compounds 1 4 ceased to increase while peak heights of analytes 5 and 6 almost retained invariable. When the injection time exceeded 150 s, peaks of analytes 3 6 became broadened and splitted. Further increasing injection time, peaks of the six analytes all became broad and splitted. The relation between injection time and peak heights was shown in Fig. 3. One probable explanation is that when a capillary is filled with sample to a fraction greater than the amount of each analyte that can be stacked, before analytes migrate toward the anode, micelles will therefore be present in the sample matrix zone, which caused broad peaks or splitted peaks. Quirino and Terabe have observed similar phenomena [12], but they did not give positive explanations. And also, this could be explained by the partial laminar flow produced inside the capillary as a result of mismatch of EOF velocity in the sample and buffer zones when the sample zone is too long [50,51]. This may also have some relationship with the retention factors of the analytes. The higher the retention factors, the faster they migrated toward the detector and interacted with reverse migration micelles, thus effectively stacked. Analytes with lower retention factors moved slower and the stacked sample zones were relatively longer, the fast migrated micelles may penetrate into them, so the broad and splitted peaks formed when injection times were long. Finally, an injection time of

6 196 J. Zhu et al. / J. Chromatogr. A 1166 (2007) Fig. 4. (A) Effect of SDS concentration on SE height. Buffer: 20 mm H 3 PO 4, mm SDS and 20% (v/v) ACN. The sample was injected using 10 kv for 10 s after introduction of 2 s water plug. (B) Effect of the injection time of water plug on the peak height. The sample was injected using 10 kv for 10 s after introduction of 0 8 s water plug. (C) Effect of electrokinetic injection time of sample on the peak height of the analytes. The sample was electrokinetically injected by 10 kv for 5 25 s after introduction of 2-s water plug. (D) Effect of the injection voltage on the peak height of the analytes. The sample was injected for 15 s using 6 to 12 kv after introduction of 2-s water plug. Buffer and other conditions are the same as those in Fig. 2C. Concentration of analytes in A, C and B, D were in the same as in Fig. 2B and C, respectively. 100 s was accepted as the optimum. Fig. 2B showed an electropherogram of the analytes with this preconcentration technique when the analytes were loaded for 100 s. Comparing the electropherograms obtained in Fig. 2A and B, the differences in migration times were probably due to the time taken to remove the sample matrix from capillary ASIW sweep RMM For this preconcentration technique, the effect of SDS concentration on SE height, length of the water plug, sample injection time, and the sample injection voltage were optimized. First, the effect of the SDS concentration on SE height was optimised. As expected, SE height increased with the increasing SDS concentration from 60 to 140 mm as shown in Fig. 4A. This may be due to the increases of the retention factors of the analytes, and the difference between the electric field strength in sample zone and that in buffer zone with the increase of SDS concentration, which will increase the stacking efficiency [44]. Though SE height was the highest at 140 mm SDS, some of the analytes were not baseline separated. Therefore, as a compromise, 100 mm SDS was selected. Next, the hydrodynamic injection time or the length of water plug was studied. The influence of the injection time of water plug on peak heights was shown in Fig. 4B. With the injection time of water plug increasing from 0 to 2 s while keeping the electrokinetic injection time at 10 s, the peak heights increased. The reason is that the water plug provides a higher electric field at the tip of the capillary, which will eventually improve the sample stacking procedure [52,53]. When the injection time was longer than 2 s, the peak heights began to decrease. This may be caused by a strong laminar flow generated due to the mismatch of EOF velocity in the sample and buffer zones [51,54]. Finally, the injection time of water plug was selected as 2 s. Third, the effect of the sample injection time was also investigated. It can be seen from Fig. 4C that the peak heights of the analytes increased with the injection time increasing from 5 to 15 s. With further increase in the injection time, the peak heights of 1 4 began to decrease. At 20 s, the peaks shape of 1 4 became gradually broad and the tips of these peaks became flat which may be due to sample overloading. The quantity of analytes injected into the capillary and the peak shape may have relationships with their electrophoretic mobilities and retention

7 J. Zhu et al. / J. Chromatogr. A 1166 (2007) factors. The larger the electrophoretic mobilities and retention factors, the faster they migrated and the stronger they interacted with micelles which move faster than suppressed EOF, thus the more can be injected and effectively swept. The electrophoretic mobilities and retention factors of compounds 1 4 were larger than that of 5 and 6, so more amounts of the four compounds were injected under the same injection time. When the amount of analyte injected reached to the maximum of the capillary that can be loaded for this analyte, the largest stacking efficiency was obtained. Further increasing sample amounts would lead to sample overloading and peak broadening. Therefore, the injection time was chosen as 15 s. Finally, effect of the sample injection voltage on the peak height was studied. The peak heights of the analytes increased with the injection voltage increasing from 6 to 10 kv as it was shown in Fig. 4D. However, further increase in the injection voltage led to the decrease of the peak heights of 5 and 6, and the peak of 4 became splitted at the bottom. This may be because that higher voltage makes the analytes get through the boundary between the water plug and the buffer and disperse into the buffer during injection procedure, which results in the decrease of the stacking efficiency [55]. Therefore, 10 kv was chosen. Under the optimum separation and stacking conditions, the electropherogram of the standard mixture of the six flavonoids was shown in Fig. 2C. The little different migration times between Fig. 2A and C were due to the time taken to remove the sample matrix from the capillary brought by different injection times. As it can be observed from Figs. 2C and 2B, the migration times were different, too. This phenomenon could be explained because the two techniques adopted two different sample injection modes: hydrodynamically and electrokinetically, respectively. In SRMM, when sample was hydrodynamically injected, the sample matrix was removed from the capillary until the EOF and the electrophoretic mobilities of the micelles analytes were balanced [46]. In contrast, in ASIW sweep RMM, electrokinetical sample injection mode was used. During injection, part of the sample matrix was pumped out of capillary already. Furthermore, electrokinetical injection itself had certain selectivity, fewer neutral sample matrixes was injected into the capillary [56]. So were the different migration times Linearity, reproducibility, LODs, and stacking enhancement factors of the two techniques Standard solutions were preconcentrated by each on-column preconcentration technique in order to determine the linearity of each compound, the reproducibility, the peak efficiency, the sensitivity enhancement, and the LODs. The results were shown in Tables 1 and 2, respectively. The results indicated that the calibration curves (based on the peak heights) were linear (r ) over the studied concentration range for SRMM and ASIW sweep RMM, respectively. The reproducibility, expressed by relative standard deviations (RSDs) of the migration times, corrected peak areas and the peak heights were studied at four different days (intra-day) and with five consecutive repeated injections at the same day (inter-day). As can be seen in Tables 1 and 2, the RSDs of the migration times, peak heights and corrected peak areas were below 3.3, 7.9, 4.8 (intra-day, n = 5), and 2.5, 12, 16 (inter-day, n = 4), respectively. There was not so much obvious difference in RSDs between the two techniques. The reproducibility of SRMM was a little better than that of ASIW sweep RMM. Table 1 Linearity, calibration graphs, reproducibility (%RSD) by migration times, corrected peak areas, and peak heights, peak efficiency, enhancement sensitivity and LODs for standard samples by SRMM Linearity ( g/ml) Calibration curves a (n =5) y = 643x y = 776x y = 937x y = 723x y = 571x 17.2 y = 316x Correlation coefficient Reproducibility (%RSD) b Intra-day (n =5) Migration time Peak height Peak area Inter-day (n =4) Migration time Peak height Peak area Peak efficiency (N) ( 10 4 ) Sensitivity enhancement SE height SE area SE LOD LOD c ( 10 3 g/ml) a y and x are the peak heights (mv) and the concentrations ( g/ml) of the analytes, respectively. b The concentrations of analytes used were as in Fig. 2B. c The detection limit was defined as the concentration where the signal-to-noise ratio is 3.

8 198 J. Zhu et al. / J. Chromatogr. A 1166 (2007) Table 2 Linearity, calibration graphs, reproducibility (%RSD) by migration times, corrected peak areas, and peak heights, peak efficiency, enhancement sensitivity and LODs for standard samples by FESI-RMM Linearity ( g/ml) Calibration curves a (n =6) y = x y = x y = x 327 y = 992x y = 893x y = 620x 106 Correlation coefficient Reproducibility (%RSD) b Intra-day (n =5) Migration time Peak height Peak area Inter-day (n =4) Migration time Peak height Peak area Peak efficiency (N) ( 10 4 ) Sensitivity enhancement SE height SE area SE LOD LOD c ( 10 3 g/ml) a As in Table 1. b The concentrations of analytes used were as in Fig. 2C. c As in Table 1. In order to compare the effectiveness of the two preconcentration techniques, the enhancement of sensitivity was investigated and compared. The sensitivity enhancement factors were evaluated by SE area,se height, and SE LOD, respectively. The results for each preconcentration technique were shown in Tables 1 and 2, respectively. Through the SE area values it can be concluded that the sensitivity was improved and fold by SRMM and ASIW sweep RMM, respectively. As can be observed, ASIW sweep RMM was the preconcentration technique that offered the higher enhancement in sensitivity when this was evaluated by SE area in comparison with SRMM technique. However, in some cases the peaks can be broad and in that case even though SE area was high, the detection capability can be low. So, SE height and SE LOD were also given in Tables 1 and 2, these values can give a better comprehension than SE area about the detection capability. Therefore, the best results of sensitivity were increased up to 16- and 31-fold by SRMM and ASIW sweep RMM, see Tables 1 and 2, respectively. Comparing to SRMM, ASIW sweep RMM was the preconcentration technique that offered the higher enhancement sensitivity when this was evaluated by SE height. The data of the peak efficiency (N) intables 1 and 2 also proved. This was because this technique adopted electrokinetical sample injection, it can greatly improve sensitivity comparing to hydrodynamical injection mode. The reason may be that analytes are injected into capillary by electrophoresis when electrokinetical injection mode is used. So, there is no injection volume restriction compared to hydrodynamical injection. Furthermore, electrokinetical injection itself can obtain certain enrichment effect in advance [57]. From Tables 1 and 2 and Fig. 1, it could be seen that the enrichment factors, SE height and SE area mainly depended on the structures of analytes. For compounds that have hydroxyls on the ring of phenyls (analytes 3 6), the stacking factors are higher when the hydroxyls were substituted with glucose than their corresponding not substituted ones. And also, for the compounds that have similar structure skeletons (1 and 2, 3 and 4, 5 and 6), the stacking factors are higher when the hydrogens on the phenyls were substituted with hydroxymethyls. Alternatively, SE LOD was also calculated as a comparison for the two preconcentration techniques. The sensitivity evaluated by SE LOD was enhanced and fold by SRMM and ASIW sweep RMM, respectively, compared to hydrodynamic injection (5 s). For the two preconcentration techniques, the LODs were calculated using a signal (height)-to-noise criterion of the three. Tables 1 and 2 also showed the LODs obtained by the two techniques for standard solutions, respectively. As can be observed, when analytes were preconcentrated by ASIW sweep RMM, the LODs were lower than for SRMM. In conclusion, it can be pointed out that the two preconcentration techniques all provided good results for the analysis of the six flavonoids compared to conventional hydrodynamic injection (5 s). By the results reported in Tables 1 and 2, it can be concluded that ASIW sweep RMM was the on-column preconcentration technique that offered better values of enhancement sensitivity in SE area, SE height, and SE LOD though the reproducibility of this technique was a little lower, and operation of this procedure was a little more complicated compared to SRMM. However, the procedure of SRMM was simpler by contrast with ASIW sweep RMM, but the injection time was longer and the enrichment factors were lower than that

9 J. Zhu et al. / J. Chromatogr. A 1166 (2007) Table 3 Contents of the analytes in Fructus aurantii Immaturus and RSD (n =4) Content (mg/g) RSD (%) of ASIW sweep RMM. Considering the sensitivity enhancement in CE with UV detection which was the main aim of this study, ASIW sweep RMM technique was selected to analyze and determine the analytes in the sample. 4. Application Fig. 5. Electropherogram of the six flavonoids in extract of Fructus aurantii Immaturus with ASIW sweep RMM by (A) 20-fold diluting; (B) 800-fold diluting of the stock sample solution. The other CE conditions were as in Fig. 2C. For experimental conditions, see text. The optimum condition of ASIW sweep RMM stacking technique was applied to the determination of the six flavonoids in the extract of F. aurantii Immaturus. The electropherogram was shown in Fig. 5. The peaks were identified by comparing the migration times and spiking the six standards individually into the sample solution. The analytical results together with RSD (n = 4) were summarized in Table 3. The contents of individual analytes were not as the same as previous reports [39 42]. Because different parts of plants and medical preparations were determined, and the ecological environment and extraction approaches also had great impact on the contents of the constituents investigated, the amounts of flavonoids in F. aurantii Immaturus could vary significantly according to their geographical sources [40]. The recovery of the method was determined with the addition of the standards in the real sample solution, with results from 90.4% to 110%. From the recovery results shown in Table 4, it can be seen that this method was accurate, sensitive and reproducible, providing a useful quantitative method for the analyses of flavonoids in F. aurantii Immaturus. Table 4 Recoveries for six analytes in F. aurantii Immaturus Ingredient Concentration spiked ( g/ml) Concentration found ( g/ml) Recovery (%) Average (%) RSD (%) Tangeretin (1) Nobiletin (2) Hesperetin (3) Naringenin (4) Hesperidin (5) Naringin (6)

10 200 J. Zhu et al. / J. Chromatogr. A 1166 (2007) Conclusion In this study, ASIW sweep RMM and SRMM were first used to analyze six flavonoids in F. aurantii Immaturus. The results showed that ASIW sweep RMM or SRMM in combination with MEKC have been proven to be a feasible and attractive way for improving the sensitivity of the detection in CE for the determination of the six analytes. With respect to the normal 5 s hydrodynamic injection, the detection sensitivity for standard solutions was increased and fold by the area of the peaks, and and fold by the height of the peaks for SRMM and ASIW sweep RMM, respectively. These techniques further exploited the potential of online focusing methods in CE. The focusing techniques were simple to perform, using conventional instrumentation without complicated procedures. These also proved that the proposed methods were suitable for flavonoid compounds. They were hopeful to be used to analyze a larger group of herbal preparations and expected to be used for pharmaceutical and herbal quality control. The preconcentration techniques allowed filling in the gap of results obtained by CE UV with standard injection. The two methods reinforced the available data concerning preconcentration of flavonoids in F. aurantii Immaturus. Furthermore, using on-line preconcentrations in CE systems made them even more sensitive, and this might be a significant improvement in flavonoids determination. And the methods were applicable in the analysis of complex samples and they were expected to use for the analysis of the minor components in plants or environment. Further investigation on the use of other sample stacking strategies in other compounds analyzing are in progress. Acknowledgement We are grateful for financial support from the National Natural Science Foundation of China (no ). References [1] S. Terabe, K. Otsuka, K. Ichikawa, A. Tsuchiya, T. Ando, Anal. Chem. 56 (1984) 111. [2] K. Otsuka, S. Terabe, Bull. Chem. Soc. Jpn. 71 (1998) [3] J.P. Chervet, R.E.J. Van Soest, M. Ursem, J. Chromatogr. 543 (1991) 439. [4] M. Albin, P.D. Grossman, S.E. Morning, Anal. Chem. 65 (1993) 489A. [5] E.E.K. Baidoo, M.R. Clench, R.F. Smith, L.W. Tetler, J. Chromatogr. B 796 (2003) 303. [6] D. Martínez, J. Farré, F. Borrull, M. Calull, J. Ruana, A. Colom, J. Chromatogr. A 808 (1998) 229. [7] L.Y. Zhu, C.H. Tu, H.K. Lee, Anal. Chem. 73 (2001) [8] Z.Y. Liu, P. Sam, S.R. Sirimanne, P.C. McClure, J. Grainger, D.G. Patterson, J. Chromatogr. A 673 (1994) 125. [9] J. Palmer, D.S. Burgi, J.P. Landers, Anal. Chem. 74 (2002) 632. [10] J. Palmer, J.P. Landers, Anal. Chem. 72 (2000) [11] J.B. Kim, K. Otsuka, S. Terabe, J. Chromatogr. A 912 (2001) 343. [12] J.P. Quirino, S. Terabe, Anal. Chem. 70 (1998) 149. [13] J.P. Quirino, S. Terabe, J. Chromatogr. A 798 (1998) 251. [14] J.P. Quirino, S. Terabe, Anal. Chem. 70 (1998) [15] J.P. Quirino, S. Terabe, Science 282 (1998) 465. [16] Y.H. Lin, M.R. Lee, R.J. Lee, W.K. Ko, Sh.M. Wu, J. Chromatogr. A 1145 (2007) 234. [17] W. Li, D. Fries, A. Alli, A. Malik, Anal. Chem. 76 (2004) 218. [18] K. Otsuka, M. Matsumura, J.B. Kim, S. Terabe, J. Pharm. Biomed. Anal. 30 (2003) [19] P. Kuban, M. Berg, C. García, B. Karlberg, J. Chromatogr. A 912 (2001) 163. [20] Q. Li, C.W. Huie, Electrophoresis 27 (2006) [21] J.B. Kim, K. Otsuka, S. Terabe, J. Chromatogr. A 932 (2001) 129. [22] J.P. Quirino, J.B. Kim, S. Terabe, J. Chromatogr. A 965 (2002) 357. [23] Zh.X. Zhang, Y.Zh. He, Y.Y. Hu, J. Chromatogr. A 1109 (2006) 285. [24] J. Musijowski, E. Pobozy, M. Trojanowicz, J. Chromatogr. A 1104 (2006) 337. [25] J.P. Quirino, S. Terabe, J. Chromatogr. A 856 (1999) 465. [26] J.B. Kim, S. Terabe, J. Pharm. Biomed. Anal. 30 (2003) [27] J.P. Quirino, S. Terabe, J. High Resolut. Chromatogr. 22 (1999) 367. [28] J. George, Clin. Chim. Acta 335 (2003) 39. [29] H. Wang, G. Cao, R.L. Prior, J. Agric. Food Chem. 45 (1997) 304. [30] B. Berkarda, H. Koyuncu, G. Soybir, Res. Exp. Med. 198 (1998) 93. [31] F.V. So, N. Guthrie, A.F. Chambers, M. Moussa, K.K. Carroll, Nutr. Cancer 26 (1996) 167. [32] T. Tanaka, H. Makita, K. Kawabata, H. Mori, M. Kakumoto, K. Satoh, Carcinogenesis 18 (1997) 957. [33] M.E. Crespo, J. Galvez, T. Cruz, M.A. Ocete, A. Zarzuelo, Planta Med. 6 (1999) 651. [34] J.A. Manthey, Microcirculation 7 (2000) S29. [35] A.D. Caro, A. Piga, V. Vacca, M. Agabbio, Food Chem. 84 (2004) 99. [36] Q. Jia, Y. Bai, Y. Ma, W. Peng, W.W. Su, Chin. Tradit. Herb. Drugs 36 (2005) 169. [37] F.I. Kanaze, C. Gabrieli, E. Kokkalou, M. Georgarakis, I. Niopas, J. Pharm. Biomed. Anal. 33 (2003) 243. [38] F. Pellati, S. Benvenuti, M. Melegari, Phytochem. Anal. 15 (2004) 220. [39] T. Wu, Y.Q. Guan, J.N. Ye, Food Chem. 100 (2007) [40] Y.Y. Peng, F.H. Liu, J.N. Ye, J. Chromatogr. B 830 (2006) 224. [41] K. Kanistr, L. Arce, A. Ríos, M. Valcárcel, Electrophoresis 22 (2001) [42] C.L. Ng, C.P. Ong, H.K. Lee, S.F.Y. Li, Chromatographia 34 (1992) 166. [43] Usio Sankawa Yui Tochun. Oversea Medicine (fascicule of Chinese Traditional Medicine), 9 (1987) 164. [44] Sh.F. Wang, Y.Q. Wu, Y. Ju, X.G. Chen, W.J. Zheng, Zh.D. Hu, J. Chromatogr. A 1017 (2003) 27. [45] J.P. Quirino, N. Inoue, S. Terabe, J. Chromatogr. A 892 (2000) 187. [46] A. Macìa, F. Borrull, M. Calull, C. Aguilar, J. Chromatogr. A 1117 (2006) 234. [47] G.Q. Song, Zh.L. Peng, J.M. Lin, J. Sep. Sci. 29 (2006) [48] A.J. García, G. Font, Y. Picó, J. Chromatogr. A 1153 (2007) 104. [49] J.F. Palmer, J. Chromatogr. A 1036 (2004) 95. [50] D.S. Burgi, R.L. Chien, Anal. Chem. 63 (1991) [51] R.L. Chien, J.C. Helmer, Anal. Chem. 63 (1991) [52] R.L. Chien, Anal. Chem. 63 (1991) [53] R.L. Chien, D.S. Burgi, J. Chromatogr. 559 (1991) 141. [54] S.H. Liu, Q.F. Li, X.G. Chen, Z.D. Hu, Electrophoresis 23 (2002) [55] A.B. Wey, W. Thormann, Chromatographia 49 (1999) S12. [56] G.B. Li, W.R. Jin, Chin. J. Anal. Instrum. 3 (1993) 1. [57] H. Xu, Y. Zhu, X.D. Yu, H.Y. Chen, Chin. J. Prog. Chem. 17 (2005) 377.