Manuel MolinaManuel Silva

Analytical Chemistry Division,Department of AnalyticalChemistry and Ecology,Campus of Rabanales,University of Córdoba,Córdoba, Spain

Simultaneous determination of phosphorus-containing amino acid-herbicides by nonionicsurfactant micellar electrokinetic chromatographywith laser-induced fluorescence detection

The potential of micellar electrokinetic chromatography (MEKC) with laser-inducedfluorescence (LIF) detection for the separation and determination of phosphorus-con-taining amino acid-herbicides (glufosinate and glyphosate), and aminomethylpho-sphonic acid (the major metabolite of glyphosate), involving derivatization with fluores-cein isothiocyanate (FITC) isomer I, was investigated. Different variables that affectderivatization (pH, FITC concentration, time and temperature) and separation (pH andconcentration of the buffer, kind and concentration of surfactants and applied voltage)were studied. The analysis was conducted within about 8 min and the use of the non-ionic surfactant Triton X-100 improved the selectivity, thus indirectly enhancing sensi-tivity by shifting of the interfering peaks of the FITC excess. Dynamic ranges of 2.0–3000 �g/L, limits of detection at microgram or submicrogram-per-liter level, and rela-tive standard deviations from 4.7 to 6.4% were obtained. The ensuing method – non-ionic surfactant MEKC – is a useful choice for the determination of these herbicidesas it provides limits of detection similar or lower than those reported by existing chro-matographic alternatives without the use of an additional preconcentration techniquesuch as solid-phase extraction. The separation of a mixture of nine FITC-derivatizedamino acids, selected as target compounds, was also carried out to assess the dis-crimination power of the nonionic surfactant MEKC method for the analysis of closelyrelated anionic analytes.

Keywords: Micellar electrokinetic chromatography / Laser-induced fluorescence / Fluoresceinisothiocyanate / Nonionic surfactants / Glyphosate / Glufosinate / Aminomethylphosphonic acid

EL 4353

1 Introduction

Glyphosate (N-(phosphonomethyl)glycine) and glufosi-nate (DL-homoalanine-4-yl-(methyl) phosphonic acid) arenonselective postemergence herbicides employed forcontrol of many grasses and broad-leafed weeds [1–8].Aminomethylphosphonic acid (AMPA) has been reportedto be the principal metabolite of glyphosate in plants,soils and waters, which is formed by microbial action [9].Once applied to vegetation, these compounds are rapidlyabsorbed by foliage and translocated throughout theplant tissue, and the residue may enter into aquatic envi-ronments [10–15], due to their high solubility in water, or

nearly completely be adsorbed by soil [16–19], in whichthe transformation of glyphosate into N-nitrosamine car-cinogen species is feasible in nitrite-treated soils [20].There is a great interest and need to determine these pes-ticides and degradation products in water, soil and plantsamples. However, it is well recognized that their chemi-cal properties (polar nature and high water solubility)which contribute to their effectiveness as herbicides alsomake their analysis difficult. On the other hand, the lack ofa chromophore or fluorophore in the molecules requiresderivatization techniques for their determination.

A number of analytical methods for the determination ofthese herbicides and their degradation products in var-ious matrices have been developed. Among chromato-graphic methods, HPLC is the preferred choice latelybecause the above-mentioned chemical properties pre-clude their analysis by GC [3, 6, 7]. HPLC with pre- orpostcolumn derivatization using 9-fluorenylmethyl chloro-formate [1, 17, 18] or o-phathalaldehyde [13, 21], respec-tively, as label reagents are the most interesting alterna-

Correspondence: Dr. Manuel Silva, Analytical Chemistry Divi-sion, Department of Analytical Chemistry And Ecology, Campusof Rabanales, University of Córdoba, E-14071 Córdoba, SpainE-mail: qa1sirom@uco.esFax: +34-957-218614

Abbreviations: AMPA, aminomethylphosphonic acid; TTAB,N-tetradecyl-N,N,N-trimenthylammonium bromide

Electrophoresis 2001, 22, 1175–1181 1175

WILEY-VCH Verlag GmbH, 69451 Weinheim, 2001 0173-0835/01/0606–1175 $17.50+.50/0

CEandCEC

tives so far. Recently, CE has emerged as a powerfulchoice to HPLC for the separation and determination ofpesticides due to its higher efficiency, separation expedi-tiousness and flexibility in selectivity adjustment [22–24].To our knowledge, only three papers have so far beendevoted to the determination of these herbicides by CE[25–27] concerning only the separation of glyphosate orglufosinate from its metabolite and no reference to multi-residue analysis of these phosphorus-containing aminoacid-herbicides appears to exist. Thus, the analysis ofglyphosate and its major metabolite was performed inwater by indirect UV detection at 240 nm under reversedpolarity conditions provided by tetradecyltrimethylammo-nium bromide [25] and in serum after derivatization withp-toluenesulfonyl chloride [26] also with UV detection.On the other hand, glufosinate and its metabolite wereseparated by using permanent coated capillaries withreversed EOF [27]. Further developments are neededbecause the reported methods lack of the suitable sensi-tivity and even selectivity to ensure efficient multiresidueanalysis of these herbicides.

In this work, a method for the sensitive determination ofglyphosate and its major metabolite (AMPA) and glufosi-nate by CE with fluorescence detection was developed.LIF detection was selected on account that lasers arethe best alternative to continuous excitation sources dueto their high intensity and collimation power, which allowsfor easy focusing of the beam into the internal diameter ofthe capillary [28]. As these analytes have no native fluor-escence, they were derivatized with fluorescein isothio-cyanate (FITC) isomer I, a very useful probe to obtainamine fluorescent derivatives for CE with LIF detection[29–35], which was used to derivatize these herbicidesfor the first time in the present work. On account that thebackground emission provided by the excess of FITCyielded interfering peaks in the electropherograms, a non-ionic surfactant such as Triton X-100 was used to over-come this shortcoming, and really this MEKC approachwas finally used for the separation of anionic FITC-deriva-tized herbicides. The results proved that nonionic surfac-tant MEKC is a powerful CE modality for the analysis ofclosely related anionic species, since separation caneasily be handled by changing the kind and concentrationof the surfactant added. Finally, the separation of nineamino acid FITC-derivatives that overlapped in the zonemodality also testified to the high separation power of theproposed nonionic surfactant MEKC approach. In addi-tion, the method is simpler than the only reference on theseparation of FITC-derivatized amino acids by MEKCreported so far by Lalljie and Sandra [33], which requiresthe use of mixed micelles (SDS/Brij-25), the addition oforganic modifiers and an analysis temperature of 35�C inorder to assure a suitable separation.

2 Materials and methods

2.1 Reagents

All chemicals used were of analytical-reagent grade andmilIi-Q water was used throughout. FITC, glyphosate, glu-fosinate and AMPA, in purity higher than 98%, were pur-chased from Riedel-de Haen (Seelze, Germany). Stan-dard solutions containing 500 �g/mL of each analytewere prepared in carbonate buffer (0.1 M sodium hydro-gen carbonate adjusted to pH 9.0 with 1.0 M sodiumhydroxide), stored at 4�C and used after dilution withmilli-Q water to the required concentration. Derivatizationwas performed using freshly prepared 2 mM FITC inchromatographic-grade acetone (Merck, Darmstadt,Germany). Other solvents were purchased from RomilChemicals (Cambridge, UK). N-Cetyl-N,N,N-trimethyl-ammonium bromide (CTAB), N-tetradecyl-N,N,N-trime-thylammonium bromide (TTAB), sodium dodecyl sulfate(SDS), Brij-35 and Triton X-100 were obtained fromMerck.

2.2 Derivatization procedure

To 400 �L of the standard solution containing a mixture ofanalytes at a concentration between 0.5 and 5000 �g/L ofeach, 100 �L of the derivatization buffer (0.65 M sodiumhydrogen carbonate adjusted to pH 8.7 with 5.0 M sodiumhydroxide) and 500 �L of 2 mM FITC acetone solutionwere added in a 1.5 mL PTFE vial. The mixture vial wascapped, homogenized and allowed to stand in the darkat 40�C for 5 h. After the derivatization, the vials werekept at –20�C till the analyses were conducted. BeforeCE analysis, the derivatization mixture was 50-fold dilutedwith milli-Q water prior to be injected into the capillaries.

2.3 MEKC

Capillary electrophoretic separations were performed ona Beckman P/ACE 5500 CE system (Beckman, Fullerton,CA, USA) equipped with a 3 mW, 488 nm air-cooledargon-ion laser (Beckman Laser Module 488). Fused-silica capillaries (50 �m ID, 363 �m OD, total length57 cm, effective length 50 cm) from Supelco (Bellefonte,PA, USA) were accommodated in a cartridge configuredfor LIF detection. A good reproducibility was assured rin-sing the capillary with 0.1 M NaOH for 5 min, followed by2 min with milli-Q water at the beginning of each experi-mental session. The capillary was equilibrated with therunning buffer (20 mM boric acid and 50 mM Triton X-100adjusted to pH 9.5 with 1.0 M sodium hydroxide) for 1 minbefore each sample injection, rinsed with 0.1 M sodiumhydroxide for 2 min and flushed with milli-Q water for1 min after each run. Samples were introduced into the

1176 M. Molina and M. Silva Electrophoresis 2001, 22, 1175–1181

Figure 1. Structure of the phoshorus-containing aminoacid-herbicides assayed (glyphosate and glufosinate)and AMPA (the major metabolite of glyphosate) and deri-vatization reaction scheme with FITC.

capillary by hydrodynamic injection, using a pressure of0.5 psi (3.45 kPa) for 10 s. Separations were performedat 25.0 ± 0.1�C using a separation voltage of 25 kV andon-line monitored with the LIF detection system (excita-tion wavelength of 488 nm and emission wavelength filterof 520 nm). All operations of the P/ACE unit were con-trolled by a PC-Pentium 75 MHz compatible computerrunning Beckman Gold Software.

3 Results and discussion

The structures and common names of glyphosate, glufo-sinate, AMPA (the metabolite of glyphosate) and FITC areshown in Fig. 1. As can be seen, the analytes possessactive hydrogen atoms in their amino group, which reactwith the isothiocyanate group of FITC to form the FITCderivatives. The migration times and peak heights of thereaction products have been used, respectively, for iden-tification and quantification of the analytes.

3.1 Optimization of the derivatizationconditions

On account that the preparation of FITC-derivatized her-bicides have not been reported in the literature and forthe subsequent optimization, we selected as initial condi-tions for this study those reported for the derivatization ofrelated compounds, such as amino acids [33, 34]. A typi-cal procedure consisted of mixing 10 �L of a standardsolution containing the analytes at a concentration of200 mg/L with 490 �L of the derivatization buffer (0.1 M

carbonate, pH 9) and 500 �L of 0.5 mM FITC in acetone.The solution was then heated at 40�C for 4 h in the dark,and after cooling, a 50-fold dilution of the derivatizationsolution was injected into the capillary for 2 s and sepa-rated at a running voltage of 25 kV. The separation bufferwas 20 mM boric acid adjusted to pH 9.5 with 0.1 M NaOH.

The effect of the pH and the concentration of the derivati-zation buffer was studied over the range pH 8.5–10.5 and0.05–0.5 M respectively. Higher pH values were not testedbecause the hydrolysis of FITC is base-catalyzed, andthus it becomes more significant at these pHs. No signifi-cant improvement of the derivatization efficiency wasobserved for variations of the pH and concentration ofthe buffer in the assayed ranges, and therefore the carbo-nate concentration and the pH of the derivatization bufferwas adjusted to 0.13 M and pH 8.9, respectively (490 �L ofthis buffer were added to reaction mixture). The pH wasfixed at 8.9 to minimize the hydrolysis of FITC and subse-quently the number of fluorescent hydrolysis products inthe FITC background emission, which allowed the detec-tion of fluorescent-herbicide derivatives at lower concen-trations.

The FITC concentration used for the derivatization wasstudied from 0.1 to 5.0 mM (500 �L of the solution in acet-one were added to the reaction mixture in all cases). Aconcentration of 2 mM was chosen as a compromisebetween maximum analytical signals and minimum inter-ference from the FITC background emission. Kineticcurves for the derivatization of the compounds studiedare illustrated in Fig. 2. At an incubation temperature of40�C the highest yields of the derivatives were attainedin ca. 5 h. These conditions were chosen as optimal,although the reaction may also be carried out at 50�C foronly 1 h with a derivatization efficiency of ca. 60% if adecreasing of the analysis time is required. In view of thissituation, further experiments were conducted to reducethe derivatization time, e.g., by using surfactants as cata-lyst. Thus, anionic (SDS), cationic (CTAB) and nonionic(Brij-35) surfactants were added to the derivatization buf-fer at concentrations higher and lower than its CMC. Theuse of these surfactants failed to increase the rate of thederivatization reaction, on the contrary, as the CMC wasreached, the reaction rate decreased, which can beascribed to an interaction of the micelles with the FITC orthe analytes.

Finally, some experiments were carried out to minimizethe sample dilution in the derivatization process. Theresults revealed that a mixture containing 100 �L of a0.65 M carbonate buffer adjusted to pH 8.7 and 400 �Lof the aqueous sample solution derivatized with 500 �Lof 0.2 mM FITC in acetone provided optimal analyticalsignals.

Electrophoresis 2001, 22, 1175–1181 Determination of herbicides by MEKC 1177

Figure 2. Kinetic curves for the derivatization of (A) glyphosate, (B) glufosinate and (C) AMPA withFITC at different temperatures: (�)30�C, (�) 40�C and (�) 50�C. Derivatization: each analyte at4 �g/mL. Samples were 25-fold diluted before injection (2 s, 0.5 psi). All other conditions as inSection 2.3.

3.2 Improvement of the separation conditionsto enhance sensitivity

The separation buffer used offered a good resolution ofthe FITC-derivatized herbicides; however, at low analyteconcentrations (below 1.0 �g/mL) it is very difficult todetermine whether the herbicides have been derivatizedor not (specially the glufosinate derivative) by comparisonto a blank, because the fluorescent hydrolysis productsand/or impurities of FITC can mask any signal attributableto a fluorescent derivative. Thus, further improvementsare needed to avoid the interference of the reagent peaksin the quantitation of the herbicides.

For our FITC herbicide mixtures, variation of the pH(8–10), temperature (15–40�C) and the concentration ofthe buffering species (10–50 mM) failed to effect thisseparation. The addition of organic modifiers, such asmethanol, ethanol, 2-propanol, n-butyl alcohol, isobutylalcohol, acetone, acetonitrile and dimethylformamideover a wide range of concentrations (5–40% dependingof the modifier used) was also ineffective in resolving thisproblem. Finally, the addition of surfactants to the runningbuffer (MEKC) was also assayed. As FITC derivatives areanionic, CTAB and TTAB at concentrations from 0.5 to5 mM were added to the buffer and the separation wascarried out under reversed polarity conditions. The inter-action between the cationic micelles and the derivativeswas so strong that the analytes coeluted as a single wideband together with the excess of FITC. Concentrations ofthe cationic surfactants below 0.5 mM caused irreprodu-cible results as the EOF can not be fully controlled. Theuse of SDS micelles was also ineffective because the

analysis time was markedly increased at concentrationsrequired to observe some difference in the migrationpattern (ca. 80 mM). In fact, the electrostatic repulsionbetween the anionic analytes and the SDS micelles inhib-ited the interaction at low SDS concentration.

Brij-35 and Triton X-100 were assayed as neutral surfac-tants at concentrations up to 80 mM. The electrophoreticcurrent was not affected by the presence of these surfac-tants and the separation showed a good reproducibility(RSD for migration time below 0.8%). As shown in Fig. 3,nonionic surfactants proved to be the most suitableapproach to minimize the effect of the FITC backgroundemission as the interaction with these peaks decreasedtheir migration times while the migration of the FITC deri-vatives was not affected. This fact can be ascribed to thepresence of the phosphonic group in the FITC derivatives,which avoids their interaction with the neutral micelles.However, this charged tail on the neutral core of the fluor-escein moiety does not appear in the hydrolysis productsof FITC, which offer a hydrophobic site to interact with theneutral micelles. Considering that the micelles migratewith the EOF, this interaction reduces the counter-EOFmobility of these compounds, so decreasing their migra-tion time.

Although both surfactants are useful to solve the problem,Triton X-100 was chosen as optimum as it interacts effi-ciently with the interfering peaks at low concentrations(down to 20 mM). A surfactant concentration of 50 mM inthe running buffer was selected to ensure a good repro-ducibility of the nonionic surfactant MEKC method withan excellent selectivity for the separation of the analytes.

1178 M. Molina and M. Silva Electrophoresis 2001, 22, 1175–1181

Figure 3. Electropherograms of a standard mixture ofanalytes at 4.0 �g/mL derivatized with FITC obtained by(A) CZE (running buffer: 20 mM sodium borate, pH 9.5) and(B) nonionic surfactant MEKC (running buffer as in (A) plus20 mM Triton X-100). Dotted and solid lines denoteblank and FITC-derivatized herbicides, respectively.Peak assignment: 1, glufosinate; 2, AMPA ; 3, glyphosate.All other conditions as in Section 2.3.

It must be observed that these experiments were per-formed with short injection times, 2 s; however, it is wellknown that a further reduction in the LOD could beachieved by increasing sample loading. A study of theinjection time up to 30 s revealed that 10 s proved to bethe optimal to avoid peak broadening while supportingthe highest sensitivity.

3.3 Analytical performance characteristics

Calibration curves for standards of herbicides and AMPAwere constructed by plotting the peak height against theanalyte concentration. The results obtained are listed inTable 1, which are referred to the volume of sample solu-tion (400 �L) subjected to the derivatization process. Allcompounds fitted a straight line with a correlation coeffi-cient of 0.997–0.999. LODs, calculated on the basis of theminimum analyte concentration that provided a chroma-tographic signal three times higher than peak-to-peaknoise, are also given in Table 1. It should be noted thatthere is no general concensus in the literature as regardsthe expression of LODs in LIF detection. As stated by Lall-jie and Sandra [34], instrumental and derivatization sensi-tivity should be considered. In the former, a relatively highconcentration of analyte is derivatized and then severaldilutions are made to calculate the LOD, whereas in thelatter, the LOD corresponds to the minimum amount/con-centration of analyte in the sample that can be deriva-tized, an additional dilution step being required to mea-sure accurately the LIF signal. Although instrumentalsensitivity affords lowest LODs (see Table 1), it is notappropriate to evaluate the potential of the developedmethod in routine analyses, and therefore derivatizationsensitivity should be used. The within-day precision(repeatibility) for the migration times and peak heights,

Table 1. Features of the calibration plots and analytical figures of merit for the determination of phosphorus-containingamino acid-herbicides by nonionic surfactant MEKC

Compond Linear range(�g/L)

Regression equationa)

(n = 14)Correlationcoefficient

LOD(�g/L)

RSD(%)

Instru-mental

Derivati-zation

200 �g/L(n = 11)

50 �g/L(n = 11)

Glufosinate 4.3–5000 H = –5.1�10–2 + 1.54�10–2 C 0.998 0.01 1.3 4.7 5.2

AMPA 1.8–2500 H = 3.0�10–2 + 2.79�10–2 C 0.999 0.004 0.5 5.6 6.4

Glyphosate 19–2000 H = 2.3�10–3 + 2.06�10–3 C 0.997 0.06 7.3 5.9 6.3

a) H, peak height (relative flurescence units); C, analyte concentration (�g/L)

Electrophoresis 2001, 22, 1175–1181 Determination of herbicides by MEKC 1179

Figure 4. Separation of a mixture of nine FITC-derivatizedamino acids by (A) CZE and (B) nonionic surfactantMEKC. Running buffer: (A) 20 mM sodium borate, pH 9.5,(B) as in (A) plus 50 mM Triton X-100. Dotted and solidlines denote blank and FITC-derivatized amino acids,respectively. Analytes were derivatized at a concentrationof 1 �g/mL. Peak assignment: 1, phenylalanine; 2, argi-nine; 3, isoleucine; 4, leucine; 5, methionine; 6, valine; 7,histidine; 8, asparagine; 9, aspartic acid. All other condi-tions as in Section 2.3.

expressed as the relative standard deviation (RSD),was calculated from 11 standards containing 50 or200 ng/mL of each analyte in water. The precisionobtained was acceptable for all the compounds, theaverage being 0.7 and 5.7% for the migration timeand peak height, respectively.

At this point, it may be interesting to compare the analyti-cal features of the proposed method for the determinationof glyphosate, AMPA and glufosinate with existing chro-matographic alternatives. GC with MS [10, 36, 37] andelectron capture [19] detection, and HPLC with fluores-cence detection, especially by using 9-fluorenylmethylchloroformate as label reagent [17, 18, 38], provide LODsat �g/L level, although the sensitivity relies on the use ofsolid-phase extraction (SPE) as a preconcentration tech-

nique. Poorer LODs (�g/mL level) are achieved by usingion chromatography with either conductimetric [11] orUV [2] detection. As stated above, there are only threereferences on the CE determination of phosphorus-con-taining amino acid-herbicides [25–27], which allowedtheir analysis in water and serum samples at �g/mL level.In general, the reported chromatographic methods (GC,HPLC and CE) are so far addressed to the determinationof one herbicide and its metabolite, the multiresidue ana-lysis being scarcely afforded. In summary, the proposedmethod is a useful choice for the determination of phos-phorus-containing amino acid-herbicides as it provideshigher sensitivity than existing CE alternatives. The sameconclusions can be drawn regarding other chromato-graphic techniques such as HPLC and GC which affordlower or similar sensitivity, in the latter supported by theuse of an additional preconcentration technique such asSPE.

3.4 Potential of nonionic surfactant MEKCfor the separation of FITC-derivatizedamino acids

Taking into account the good features achieved with theproposed nonionic surfactant MEKC method in theseparation of FITC-derivatized herbicides, we have eval-uated its potential for the resolution of the unsolved prob-lems so far in the analysis of some FITC-derivatizedamino acids by MEKC [33, 34] due to their high structureresembling. Eight amino acids were chosen as target ana-lytes as they include three groups of amino acids thatshow important overlapping trouble: leucine/isoleucine,methionine/phenylalanine/valine and histidine/aspara-gine/arginine. Aspartic acid was also selected to coverthe three types of amino acids (neutral, acidic and basic).The amino acids were derivatized at a concentration of1 mg/L as described in Section 2.2 and then separatedunder CZE (running buffer: 20 mM borate, pH 9.5) andnonionic surfactant MEKC (the same buffer as for CZEcontaining 50 mM Triton X-100). As can be seen in Fig. 4,CZE was ineffective for the separation of these analytes,whereas the nine FITC-derivatized amino acids werecompletely separated by the nonionic surfactant MEKCmodality. These results clearly testify to the high separa-tion power of this MEKC modality for the analysis ofclosely-related anionic analytes.

4 Concluding remarks

In this paper, nonionic surfactant MEKC is shown to bea powerful analytical technique for the determination ofclosely-related anionic compounds, such as FITC-deriva-tized herbicides and amino acids. To our knowledge, FITC

1180 M. Molina and M. Silva Electrophoresis 2001, 22, 1175–1181

has for the first time been used as a fluorescent probe forphosphorus-containing amino acid-herbicides. The sim-plicity of the proposed method makes it suitable for themultifresidue analysis of these compounds in a greatvariety of samples. The following conclusions can bedrawn from its performance: (i) the method extends thescope of MEKC to the determination of phosphorus-con-taining amino acid-herbicides, for which this techniquehas scarcely been used so far. To our knowledge, onlythree recent papers on this topic have been published[25–27], neither of which was concerned with the determi-nation of a mixture of the three compounds studied in thiswork. (ii) The proposed nonionic surfactant MEKC modal-ity allows the complete separation of the analytes andinterfering peaks resulting from the excess of FITC. Theselectivity of the separation can easily be adjusted by thespecific surfactant and its concentration. Moreover, thisprocedure can be generalized to the separation of closelyrelated anionic species, as has been shown in the separa-tion of nine FITC-derivatized amino acids, (iii) The LODsobtained (at few or submicrogram-per-liter level and simi-lar or lower than those of existing chromatographic alter-natives) make the proposed method a useful choice forthe determination of the herbicides in environmentalwaters and foods, although in the former, a previousenrichment of the analytes can easily be carried outthrough ion-exchange columns if their determination atfew ng/L level is required.

The authors gratefully acknowledge financial supportfrom Spain’s Dirección General Interministerial de Cienciay Tecnologia (DGICyT) for the realization of this work aspart of Project Bau 2000-0905.

Received October 12, 2000

5 References

[1] Hogendoorn, E. A., Ossendrijver, F. M., Dijkman E., Bua-mann, R. A., J. Chromatogr. A 1999, 833, 67–73.

[2] Morlier, L. W, Tomkins, D. F, J. AOAC Int. 1997, 80, 464–468.[3] Tsuji, M., Akiyama, Y., Yano, M.,Anal. Sci. 1997,13, 283–285.[4] Sen, N. P., Baddoo, P. A., Int. J. Environ. Anal. Chem. 1996,

63, 107–117.[5] Kataoka, H., Ryu, S., Sakiyama, N., Makita, M., J. Chroma-

togr. A 1996, 726, 253–258.[6] Alferness, P. L., Iwata, Y., J. Agric. Food Chem. 1994, 42,

2751–2759.[7] Wigfield, Y. Y., Lanouette, M., J. Assoc. Off. Anal. Chem.

1991, 74, 842–847.[8] Konar, S. K., Roy, D. N.,Anal.Chim.Acta1990,229, 277–280.[9] Rueppel, M. L., Brightwell, B. B., Schaefer, J., Marvel, J. T.,

J. Agric. Food Chem. 1977, 25, 517–528.[10] Stalikas, C. D., Pilidis, G. A., Karayannis, M. I., Chromato-

graphia 2000, 51, 741–746.

[11] Zhu, Y., Zhang, F. F., Tong, C. L, Liu, W. P., J. Chromatogr. A1999, 850, 297–301.

[12] Skark, C., Zullei-Seibert, N., Schottler, U., Schlett, C., Int. J.Environ. Anal. Chem. 1998, 70, 93–104.

[13] Mallat, E., Barcelo, D., J. Chromatogr. A 1998, 823, 129–136.

[14] Mogadati, P. S., Louis, J. B., Rosen, J. D., J. AOAC Int. 1996,79, 157–162.

[15] Sancho J. V., Hernandez, F., Lopez, F. J., Hogendoom, E. A.,Dijkman, E., van-Zoonen, P., J. Chromatogr. A 1996, 737,75–83.

[16] Ravanel, P., Liegeois, M. H., Chevallier, D., Tissut, M., J.Chromatogr. A 1999, 864, 145–154.

[17] Smith, N. J., Martin, R. C., Croix, R. G., Bull. Environ. Con-tam. Toxicol. 1996, 57, 759–765.

[18] Sancho J. V., Hidalgo, C., Hernandez, F., Lopez, F. J.,Hogendoorn, E. A., Dijkman, E., Int. J. Environ. Anal. Chem.1996, 62, 53–63.

[19] Eberbach, P. L., Douglas, L. A., J. Agric. Food Chem. 1991,39, 1776–1780.

[20] Pastore, P., Lavagnini, I., Boaretto, A., Magno, F., Anal.Chim. Acta 1990, 230, 29–34.

[21] Sundaram, K. M. S., Curry, J., J. Liq. Chromatogr. Rel. Tech-nol. 1997, 20, 511–524.

[22] Karcher, A., ElRassi, Z., Electrophoresis 1999, 20, 3280–3296.

[23] Takeda, S., Chromatographia 1999, 20, 150–151.

[24] Juhler, R. K., Lauridsen, M. G., Christensen, M. R., Hilbert,G., J. AOAC Int. 1999, 82, 337–358.

[25] Cikalo, M. G., Goodall, D. M., Matthews, W., J.Chromatogr. A1996, 745, 189–200.

[26] Tomita, M., Okuyama, T., Nigo, Y., Uno, B., Kawai, S., J.Chromatogr. 1991, 109, 324–330.

[27] Finkler, C., Charrel, H, Engelhardt, H., J. Chromatogr. A1998, 822, 101–106.

[28] Swinney, K., Bornhop, D. J., Electrophoresis 2000, 21,1239–1250.

[29] Jin, L. J., Rodriguez, I., Li, S. F. Y., Electrophoresis 1999,20,139–148.

[30] Dabek-Zlotorzynska, E., Maruszak, W., J. Chromatogr. B1998, 714, 77–85.

[31] Little, M. J., Paquette, D. M., Harvey, M. D., Banks, P. R.,Anal. Chim. Acta 1997, 339, 279–288.

[32] Rodriguez, I., Lee, H. K., Li, S. F. Y., J. Chromatogr. A 1996,745, 255–262.

[33] Lalljie, S. P. D., Sandra, P., Chromatographia 1995, 40,513–518.

[34] Lalljie, S. P. D., Sandra, P., Chromatographia 1995, 40,519–526.

[35] Mattusch, J., Dittrich, K., J. Chromatogr. A 1994, 680,279–285.

[36] Bauer, K. H., Knepper, T. P., Maes, A., Schatz, V., Voihsel, M.,J. Chromatogr. A 1999, 837, 117–128.

[37] Royer, A., Beguin, S., Tabet, J. C., Hulot, S., Reding, M. A.,Communal, P. Y, Anal. Chem. 2000, 72, 3826–3832.

[38] Ohno, C., Otaki, M., Mori, Y., Hisamatsu, Y., Nakazawa, H.,Shokuhin Eiseigaku Zasshi 1999, 40, 75–79.

Electrophoresis 2001, 22, 1175–1181 Determination of herbicides by MEKC 1181