Comparative study on sample stacking by moving reaction boundary formed with weak acid and weak or strong alkali in capillary electrophoresis: I. Theory

The condensation of low abundance zwitterion substance, such as protein and peptide, has great significance to the study on proteomics. This paper develops the theory on design of online stacking conditions of zwitterion by a moving reaction boundary (MRB) in capillary electrophoresis (CE). This concerns the choice of running and sample buffers, velocity design of MRB, and salt effect on the stacking. The theoretical results unveil that: (1) the velocity of MRB formed with weak acidic buffer and strong alkali should be set between zero and the velocity of zwitterion in the alkali phase, or no stacking occurs; (2) if a strong alkali is used to prepare the sample, a much long front plug of strong base must be injected before the alkaline sample plug for complete stacking, whereas no such front plug is needed if a weak alkali with enough high concentration and pH value is used to prepare the sample buffer; (3) the existence of salt in sample matrix has a weak effect on the stacking of zwitterion if sample is prepared with weak alkaline buffer, while has a dramatic effect on the same stacking if with a strong base buffer. In addition, the concentration of weak alkali used for preparation of sample should be set at the point, at which the velocity of MRB is as much as possible close to that of negative zwitterion. The developed theory and its computation are quantitatively proved by the experiments of zwitterion stacking by the MRB as shown in the previous and the accompanying papers. The proposed theoretic results hold obvious significances on-column stacking of low abundance zwitterions, such as amino acid, or peptides or proteins, in CE.     

Comparative study on sample stacking by moving reaction boundary formed with weak acid and weak or strong base in capillary electrophoresis: II. Experiments

To demonstrate the theoretic method on the stacking of zwitterion with moving reaction boundary (MRB) in the accompanying paper, the relevant experiments were performed. The experimental results quantitatively show that (1) MRB velocity, including the comparisons between MRB and zwitterionic velocities, possesses key importance to the design of MRB stacking; (2) a much long front alkaline plug without sample should be injected before the sample injection for a complete stacking of zwitterion if sample buffer is prepared with strong base, conversely no such plug is needed if using a weak base as the sample buffer with proper concentration and pH value; (3) the presence of salt in MRB system holds dramatic effect on the MRB stacking if sample solution is a strong base, but has no effect if a weak alkali is used as sample solution; (4) all of the experiments of this paper, including the previous work, quantitatively manifest the theory and predictions shown in the accompanying paper. In addition, the so-called derivative MRB-induced re-stacking and transient FASI-induced re-stacking were also observed during the experiments, and the relevant mechanisms were briefly demonstrated with the results. The theory and its calculation procedures developed in the accompanying paper can be well used for the predictions to the MRB stacking of zwitterion in CE.     

Study on mechanism of stacking of zwitterion in highly saline biologic sample by transient moving reaction boundary created by formic buffer and conjugate base in capillary electrophoresis

The reason why a moving reaction boundary (MRB) can stack analyte in highly saline sample in capillary electrophoresis [C.X. Cao, Y.Z. He, M. Li, Y.T. Qian, S.L. Zhou, L. Yang, Q.S. Qu, Anal. Chem. 74 (2002) 4167] is still unclear. To illuminate the mechanism of such stacking, three MRBs formed by formic acid-NaOH buffer and sodium formate as well as 40, 80 and 120 mmol/L sodium chloride in matrixes were studied. The computation with MRB theory shows that sodium chloride in matrix has weak effect on the stacking efficiency, whether the concentration of sodium chloride is set at 40, or 80, or 120 mmol/L. The conclusion has been highly manifested by numerous experiments. Furthermore, the computer simulation and theoretical analyses depict that this kind of stacking is induced by the mechanism of MRB, rather than that of electrostacking or isotachophoresis (ITP) under the given electrolytic system. Finally, the application of the sample condensation was achieved for the stacking of analyte(s) in highly saline biological sample of skeletonema costarum’ culture with up to 527 mmol/L total salt and health human urine with 150-320 mmol/L inorganic ions (Cl⁻, Na⁺, K⁺, PO₄³⁻, etc.). The results herein have a clear significance to the design on stacking of analyte in highly saline biological sample.     

Sample preconcentration by field amplification stacking for microchip-based capillary electrophoresis

A microchip structure for field amplification stacking (FAS) was developed, which allowed the formation of comparatively long, volumetrically defined sample plugs with a minimal electrophoretic bias. Up to 20-fold signal gains were achieved by injection and separation of 400 microm long plugs in a 7.5 cm long channel. We studied fluidic effects arising when solutions with mismatched ionic strengths are electrokinetically handled on microchips. In particular, the generation of pressure-driven Poiseuille flow effects in the capillary system due to different electroosmotic flow velocities in adjacent solution zones could clearly be observed by video imaging. The formation of a sample plug, stacking of the analyte and subsequent release into the separation column showed that careful control of electric fields in the side channels of the injection element is essential. To further improve the signal gain, a new chip layout was developed for full-column stacking with subsequent sample matrix removal by polarity switching. The design features a coupled-column structure with separate stacking and capillary electrophoresis (CE) channels, showing signal enhancements of up to 65-fold for a 69 mm long stacking channel.     

Determination of mercaptopurine and its four metabolites by large-volume sample stacking with polarity switching in capillary electrophoresis

This study describes approaches for stacking a large volume of sample solutions containing a mixture of mercaptopurine monohydrate, 6-methylmercaptopurine, thioguanine, thioguanosine, and thioxanthine in capillary electrophoresis (CE). After filling the run buffer (60 mM borate buffer, pH 8.5), a large sample volume was loaded by hydrodynamic injection (2.5 psi, 99.9 s), followed by the removal of the large plug of sample matrix from the capillary using polarity switching (-15 kV). Monitoring the current and reversing the polarity when 95% of current recovered, the separation of anionic analytes was performed in a run buffer < 20 kV. Around 44- to 90-fold improvement of sensitivity for five analytes was achieved by large-volume stacking with polarity switching when compared with CE without stacking. This method was feasible for determination of the analytes spiked in plasma. Removing most of electrolytes from plasma is a key step for performing large-volume sample stacking. Solid-phase extraction was used for pretreatment of biological samples. To our knowledge, this study is one of few applications showing the possibilities of this stacking procedure to analyze biological samples by large-volume sample stacking with polarity switching (LVSSPS) in CE.     

Solid-phase extraction and large-volume sample stacking with an electroosmotic flow pump in capillary electrophoresis for determination of methotrexate and its metabolites in human plasma

This paper describes approaches for large-volume sample stacking (LVSS) with an EOF pumpin CE for the determination of methotrexate (MTX) and its metabolites in human plasma. After pretreatment of plasma through a SPE cartridge, a large sample volume was loaded by hydrodynamic injection (3 psi, 70 s) into the capillary filled with phosphate buffer (70 mM, pH 6.0) containing 0.01% polyethylene oxide. Following removal of a large plug of sample matrix from the capillary using polarity switching (-25 kV), the separation of anionic analytes was subsequently performed without changing polarity again, achieving an improvement of sensitivity of around a 100-fold. The method was applied to therapeutic drug monitoring of MTX in one acute lymphoblastic leukemia patient. This study is one of very few applications showing the feasibility of LVSS in analysis of biological samples by CE.     

Change of migration time and separation window accompanied by field-enhanced sample stacking in capillary zone electrophoresis.

When field-enhanced sample stacking was used in capillary zone electrophoresis (CZE) analysis of cations, the decrease of migration time and the reduction of separation window was observed with increase of sample plug length. A simple equation expressing the migration velocity in the stacking process was derived to explain the above phenomenon. From experiments and theoretical consideration, we confirmed that this effect was caused by the higher potential gradient and larger eletroosmotic flow (EOF) mobility at the sample plug than those at the supporting electrolyte. A mathematical model appropriate for the computer simulation of such a system was studied considering the experimental results, and it was concluded that electroosmotic velocity (v(eof)) should be introduced to the equation of continuity as a constant.     

A highly sensitive method for enantioseparation of fenoprofen and amino acid derivatives by capillary electrophoresis with on-line sample preconcentration

A highly sensitive method for enantioseparation of trace fenoprofen and amino acid derivatives by capillary electrophoresis (CE) with vancomycin as the chiral selector was developed. Several CE techniques, such as the partial filling, large-volume sample stacking with EOF as pump plus anion-selective exhaustive injection (LVSEP-ASEI) were involved in the present method to improve the detection sensitivity. With on-column concentration, enantioseparation of racemic fenoprofen and six 9-fluorenylmethyl chloroformate (FMOC)-amino acid derivatives (at the concentration level of ng/mL) with the background electrolyte composed of 100 mmol/L Tris-H(3)PO(4) (pH 6.0) and 2 mmol/L vancomycin was detected readily with the UV detection at 214 nm. Successfully performing LVSEP-ASEI needs a very low EOF that could be depressed by coating the capillary with poly(dimethylacrylamide) solution. The coating also played a role to minimize the adsorption of vancomycin onto the capillary wall. Effect of the injected sample volume and the electrokinetic injection time on the peak area of the enantiomers and their resolution factor were investigated and optimized. Under the optimized conditions, more than 1000-fold enhancement in detection sensitivity compared with the normal injection was achieved.     

Using multiple short-end injections to develop fast electrophoretic separations--applications in iodide analysis

The aim of this study was to develop a fast CE separation method by using multiple short-end injections in a capillary coated with quaternary ammonium chitosan (HACC), in order to determine the iodide content of pharmaceutical formulations. The BGE was composed of 20 mM tris(hydroxymethyl)aminomethane and 11 mM hydrochloric acid, at pH 8. The internal standard used was thiocyanate. Separations were performed in a fused silica capillary (32 cm total length, 8.5 cm effective length and 50 μm i.d.) coated with HACC and direct UV detection at 220 nm. EOF was modified by flushing the capillary with polymeric solution, resulting in a semi-permanent coating of controlled and stable EOF. The EOF was anodic at pH 8. Different strategies, using single and multiple injection short-end configurations, were studied to develop a CE method that resulted in a maximum number of iodide samples analyzed per hour: one plug and flush (Sflush) 35 samples/h, one plug without flush (SWflush) 76 samples/h, four plugs and flush (Mflush) 61 samples/h, and four plugs without flush (MWflush) 80 samples/h. Using the multiple injection configuration, it was possible to inject up to four plugs using spacer electrolytes with good separation efficiency and selectivity. The voltage application time needed to separate the eight peaks (iodide and thiocyanate) with MWflush was only 12s. The method was validated and samples were analyzed using MWflush. Good linearity (R(2)>0.999); a limit of detection 0.4 mg L(-1); intermediate precision better than 3.8% (peak area) and recovery in the range of 99-102% were obtained.     

Different sample stacking strategies to analyse some nonsteroidal anti-inflammatory drugs by micellar electrokinetic capillary chromatography in mineral waters

Three on-column preconcentration techniques were compared to analyse a group of nonsteroidal anti-inflammatory drugs (NSAIDs) using micellar electrokinetic capillary chromatography (MEKC) under pH-suppressed electroosmotic flow (EOF) in water samples. The analysed drugs were ibuprofen, fenoprofen, naproxen, ketoprofen, and diclofenac sodium. The micellar background electrolyte (BGE) solution was formed by 75 mM sodium dodecyl sulfate (SDS), 40% (v/v) acetonitrile, and 25 mM sodium phosphate at pH 2.5. When this BGE solution was used the applied voltage was reversed, -10 kV, and the drugs were separated within 20 min. The on-column preconcentration modes, characterised all of them for the sample matrix removal out of the capillary by itself under a reverse potential at the same time as the EOF was reduced, were stacking with reverse migrating micelles (SRMM), stacking with reverse migrating micelles-anion selective exhaustive injection (SRMM-ASEI), and field-enhanced sample injection with reverse migrating micelles (FESI-RMM). The sensitivity was improved up to 154-, 263-, and 63-fold, respectively when it was calculated through the peaks height. The optimised methods were validated with spiked mineral water by combining off-line solid-phase extraction (SPE) and the proposed on-line sample stacking strategies. The detection limits (LODs) of NSAIDs in mineral water were at ng/L levels.     

Novel moving reaction boundary-induced stacking and separation of human hemoglobins in slab polyacrylamide gel electrophoresis

We developed a novel polyacrylamide gel electrophoresis (PAGE) method to stack and separate human hemoglobins (Hbs) based on the concept of moving reaction boundary (MRB). This differs from the classic isotachophoresis (ITP)-based stacking PAGE in the aspect of buffer composition, including the electrode buffer (pH 8.62 Tris–Gly), sample buffer (pH 6.78 Tris–Gly), and separation buffer (pH 8.52 Tris–Gly). In the MRB-PAGE system, a transient MRB was formed between alkaline electrode buffer and acidic sample buffer, being designed to move toward the anode. Hbs carried partial positive charges in the sample buffer due to its pH below pI values of Hbs, resulting in electromigrating to the cathode. Hbs would carry negative charges quickly when migrated into the alkaline electrode buffer and be transported to the anode until meeting the sample buffer again. Thus, Hbs were stacked within a MRB until the transient MRB reached the separation buffer and then separated by zone electrophoresis with molecular sieve effect of the gel. The experimental results demonstrated that there were three clear and sharp protein zones of Hbs (HbA_1c, HbA₀, and HbA₂) in MRB-PAGE, in contrast to only one protein zone (HbA₀) in ITP-PAGE for large-volume loading (≥15 μl), indicating high stacking efficiency, separation resolution, and good sensitivity of MRB-PAGE. In addition, MRB-PAGE was performed in a conventional slab PAGE device, requiring no special device. Thus, it could be widely used in separation and analysis of diluted protein in a standard laboratory.

Fluorescence imaging of sample zone narrowing and dispersion in a glass microchip: the effects of organic solvent (acetonitrile)-salt mixtures in the sample matrix and surfactant micelles in the running buffer

A mismatch in the EOF velocities between the sample zone and running buffer region is known to generate pressure-driven, parabolic flow profile of the sample plug in electrokinetic separation systems. In the present study, video fluorescence microscopy was employed to capture real-time dynamics of the sample plug (containing fluorescein as the probe molecule) in a discontinuous conductivity system within a glass microchip, in which the sample matrix consisted of a mixture of ACN and salt (NaCl), and the running buffer contained sodium cholate (SC) micelles as the pseudo-stationary phase (i.e., performing "ACN stacking" in the mode of MEKC). Upon application of the separation voltage, the video images revealed that zone narrowing and broadening of the probe molecules occurred as the sample plug headed toward the cathode during the initial time period, probably resulting in part from the stacking/sweeping, and destacking of the SC micelles at the boundaries between the sample zone and running buffer. Interestingly, a second sample zone narrowing event can be observed as the sample plug moved further toward the cathode, which could be attributed to the sweeping of the slower moving probe molecules by the faster moving SC micelles that originated from the anode. This phenomenon was studied as a function of pH, sample plug length, as well as the concentration of organic solvent and salt in the sample matrix. The data suggested that the presence of large amounts of an organic solvent (such as ACN or methanol) and salts in the sample matrix not only induces sample dispersion due to the formation of a pressure-driven (hydrodynamic) flow, but may also lead to the formation of a double sample zone narrowing phenomenon by altering the local EOF dynamics within the separation system.     

Non-uniform surface charge distributions in CE: theoretical and experimental approach based on Taylor dispersion

The control of the EOF direction and magnitude remains one of the more challenging issues for the optimization of separations in CE. In this work, we investigated the possibility to use non-uniform surface charge distribution for the modulation of the EOF in CE. Non-uniform zeta potentials were obtained by modifying a section of the capillary surface using adsorption of polyelectrolytes. Three different methods were studied: (i) partial polycation coating on a fused silica capillary, (ii) partial polycation (or polyanion) coating on polyelectrolyte multilayers, and (iii) partial polycation coating on a capillary previously modified with poly(ethylene oxide). The magnitude and the direction of the EOF as a function of the coated capillary length were first studied. The stability of the EOF and the separation performances were also considered taking two dialanine diastereoisomers as model compounds. In partially coated capillaries, the average solvent flow is the sum of two contributions: a non-dispersive electroosmotic contribution related to the capillary surface charge, and a dispersive hydrodynamic contribution that depends on the difference of surface charge between the coated and the non-coated capillary zones. To get a better insight into the influence of the hydrodynamic contribution to the total peak dispersion, the peak variances corresponding to the Taylor dispersion, the injection plug, and the axial diffusion were calculated. This work demonstrates that peak dispersion in a capillary partially coated by the inlet end is different from that obtained when the coating is performed by the outlet end. Experimentally, the combination of a partially coated capillary with a large volume sample stacking preconcentration step can be used for injecting up to 95% of the capillary volume. This approach leads to a preconcentration factor of 60 compared with CZE with classical injection.     

应用移动反应界面富集技术进行毛细管电泳尿液指纹分析

快速灵敏的尿液指纹图谱分析对于临床诊断中发现新的生物标记至关重要。该文建立了一种简便、快速、灵敏的移动反应界面(MRB)介导的富集技术进行毛细管电泳尿液指纹图谱分析。MRB由25 mmol/L甘氨酸（Gly）-HCl（pH 2.5）作为样品缓冲液和50 mmol/L Gly-NaOH（pH 12.3）作为电泳缓冲液形成。与常规的毛细管区带电泳只能观察到尿液中不到10个峰相比,采用MRB可以观察到超过80个峰并将检测灵敏度提高了至少十几倍,显示该方法对于代谢组学分析具有重要的意义。     

Capillary electrophoretic analysis of carbohydrates derivatized by in-capillary condensation with 1-phenyl-3-methyl-5-pyrazolone

Our previous papers on capillary electrophoresis (CE) have shown that samples can be derivatized in a capillary and the derivatives can be analyzed immediately after derivatization, provided that the derivatization reaction is so rapid as to complete in seconds. The present paper presents extended application of in-capillary derivatization to a much slower reaction such as the condensation of reducing carbohydrates with 1-phenyl-3-methyl-5-pyrazolone (PMP) which requires 30 min at 70 degrees C in pre-column derivatization by manual operation. It was necessary to first drive the introduced plugs of sample and reagent solutions to put them together at the entrance of the heated portion of a capillary, then to allow the superimposed plugs to react for a relevant period. We showed how to determine the introduction times of the sample and the reagent solutions as well as intermediate running buffer, the voltages to be applied for plug driving and product analysis, and the duration of voltage application, all of which are important for effective in-capillary derivatization. An example of the analysis of maltooligosaccharides by this technique is presented. It was shown that maltooligosaccharides were quantitatively derivatized with PMP in 35 min at 57 degrees C, and the derivatives could be analyzed in ca. 15 min by CE immediately after derivatization. Separation was satisfactory in 200 mM borate buffer, pH 8.2 containing sodium dodecyl sulfate to a concentration of 200 mM. Although the theoretical plate number, and accordingly the resolution, were significantly lower than the corresponding values in pre-capillary derivatization, reasonable reproducibility was ensured for both migration time (RSD 3.5% on average) and peak area (RSD less than 3%) under the optimized conditions. It is notable that sample amount could be lowered to the 10 fmol level, in contrast to the 10 pmol level in pre-capillary derivatization. In addition, since the technique employed here (the modified at-inlet technique of in-capillary derivatization) is easily automated, the established system will be highly beneficial for routine analysis of carbohydrates. Analysis by this technique was also shown to be useful for kinetic study of the derivatization reaction.     

Online transient micellar phase concentration of anions using CTAB in CE

A transient micellar phase extractor using CTAB was described for the online sample concentration of various anionic analytes (drugs and herbicides) in CE. Stacking and separation was performed at neutral pH in coelectroosmotic flow in a hexadimethrine bromide coated fused‐silica capillary. A micellar plug (e.g. 10 mM CTAB) was injected prior to hydrodynamic injection of the analytes prepared in aqueous organic solvent (e.g. with 30% ACN). In the presence of an electric field, the micelles interacted with the anions inside the capillary. This was followed by selective analyte focusing via the mechanism of micelle to solvent stacking. The micelles acted as transient extractor because the stacking ends when the injected micelles completely migrated through the boundary between the sample and micellar plug. Fundamental studies were performed (effect of surfactant concentration, etc.) and the technique yielded 13‐ to 30‐fold improvements in peak height. A stacking CE method in conjunction with liquid–liquid extraction was also tested for the detection of the herbicides fenoprop and mecoprop in fortified drinking water at analyte concentration levels relevant to Australian Drinking Water Guidelines.     

Optimization of sample stacking for the simultaneous determination of nonsteroidal anti-inflammatory drugs with a wall-coated histidine capillary column

A wall-coated histidine capillary column was developed for the on-line preconcentration of nonsteroidal anti-inflammatory drugs (NSAIDs) in capillary electrochromatography (CEC). A wide variety of experimental parameters, such as the sample buffer, background electrolyte (BGE) composition, concentration, sample plug lengths, water plug, and the effect of organic modifiers were studied. The relationship between peak height and injection times for the NSAIDs by variation of sample and BGE buffer concentration was investigated. On addition of sodium chloride (0.3-0.6%) to the sample zone, the stacking efficiency was increased. With acetate buffer (100 mM, pH 5.0)/ethanol (20% v/v) as BGE and sample solution in acetate buffer (0.2 mM, pH 5.0)/ethanol (20% v/v)/NaCl (0.3% w/v), NSAIDs could be determined at low microM levels without sample matrix removal. The detection limit was 0.096 microM for indoprofen, 0.110 microM for ketoprofen, 0.012 microM for naproxen, 0.023 microM for ibuprofen, 0.110 microM for fenoprofen, 0.140 microM for flurbiprofen, and 0.120 microM for suprofen. The method could be successfully applied to the simultaneous determination of NSAIDs in urine. The recoveries were better than 82% for all the analytes. The present method enables simple manipulation with UV detection for the determination of NSAIDs at low concentration levels in complex matrix samples.     

pH-mediated acid stacking with reverse pressure for the analysis of cationic pharmaceuticals in capillary electrophoresis

When using capillary electrophoresis (CE) for the analysis of biological samples, it is often necessary to employ techniques to overcome peak-broadening that results from having a high-conductivity sample matrix. To improve the concentration detection limits and separation efficiency of cationic pharmaceuticals in CE, pH-mediated acid stacking was performed to electrofocus the sample, improving separation sensitivity for the analyzed cations by 60-fold. However, this method introduces a large titrated acid plug into the capillary. To overcome the limitations this low-conductivity plug poses to stacking, the plug was removed prior to the separation step by applying reverse pressure to force it out of the anode of the capillary. Employing this technique allows for roughly twice the volume of sample to be injected. A maximum sample injection time of 240 s was attainable with baseline peak resolution compared to a maximum sample injection time of 120 s without reverse pressure, leading to a twofold decrease in the limits of detection of the analytes used. Separation efficiency overall is also improved when utilizing the reverse pressure step. For example, a 60 s sample injection time results in 94,000 theoretical plates as compared to 60,500 theoretical plates without reverse pressure. This reverse-pressure method was used for detection and quantitation of several cationic pharmaceuticals that were prepared in Ringer's solution to simulate microdialysis sampling conditions.     

Improving sensitivity by large-volume sample stacking combined with sweeping without polarity switching by capillary electrophoresis coupled to photodiode array ultraviolet detection

An easy, simple, and highly efficient on-line preconcentration method for polyphenolic compounds in CE was developed. It combined two on-line concentration techniques, large-volume sample stacking (LVSS) and sweeping. The analytes preconcentration technique was carried out by pressure injection of large-volume sample followed by the EOF as a pump pushing the bulk of low-conductivity sample matrix out of the outlet of the capillary without the electrode polarity switching technique using five polyphenols as the model analytes. Identification and quantification of the analytes were performed by photodiode array UV (PDA) detection. The optimal BGE used for separation and preconcentration was a solution composed of 10 mM borate-90 mM sodium cholate (SC)-40% v/v ethylene glycol, without pH adjustment, the applied voltage was 27.5 kV. Under optimal preconcentration conditions (sample injection 99 s at 0.5 psi), the enhancement in the detection sensitivities of the peak height and peak area of the analytes using the on-line concentration technique was in the range of 18-26- and 23-44-fold comparing with the conventional injection mode (3 s). The detection limits for (-)-epigallocatechin (EGC), (-)-epicatechin (EC), (+)-catechin (C), (-)-epigallocatechin gallate (EGCG), and (-)-epicatechin gallate (ECG) were 4.3, 2.4, 2.2, 2.0, and 1.6 ng/mL, respectively. The five analytes were baseline-separated under the optimum conditions and the experimental results showed that preconcentration was well achieved.     

On-Column Preconcentration of Anti-Inflammatory Drugs in River Water by Anion-Selective Exhaustive Injection-Sweeping-MEKC

A high-sensitivity on-column preconcentration method, anion selective exhaustive injection (ASEI)-sweeping in micellar electrokinetic capillary chromatography (MEKC) has been developed for the analysis of non-steroidal anti-inflammatory drugs (NSAIDs): diclofenac sodium, ibuprofen, fenoprofen, naproxen, and ketoprofen in water samples. To achieve the best results of the ASEI-sweeping-MEKC method, conditions which affected preconcentration were examined, including the sodium dodecyl sulfate concentration, composition of the sample matrix, composition and injection length of the water plug, the concentration and the injection length of the high-conductivity buffer, and finally the sample injection time. Under the optimum stacking conditions the method was validated for the determination of the studied NSAIDs in river water samples with limits of detection ranging between 29 and 58 ng mL^?1, and without any previous sample pretreatment.