Dispersive liquid–liquid microextraction combined with high-performance liquid chromatography–tandem mass spectrometry for the identification and the accurate quantification by

J. Sep. Sci. 2013, 36, 3067–3074 3067

Yaohai ZhangBining Jiao

Citrus Research InstituteSouthwest University/Laboratoryof Citrus Quality and Safety RiskAssessment, Ministry ofAgriculture/Key Laboratory ofHorticulture Science for SouthernMountainous Regions of Ministryof Education, SouthwestUniversity, Chongqing, China

Received March 15, 2013Revised June 8, 2013Accepted June 13, 2013

Research Article

Dispersive liquid–liquid microextractioncombined with online preconcentrationMEKC for the determination of somephenoxyacetic acids in drinking water

A fast and simple technique composed of dispersive liquid–liquid microextraction(DLLME) and online preconcentration MEKC with diode array detection was developedfor the determination of four phenoxyacetic acids, 2,4,5-trichlorophenoxyacetic acid, 2,4-dichlorophenoxyacetic acid, 2,6-dichlorophenoxyacetic acid, and 4-chlorophenoxyacetic acid,in drinking water. The four phenoxyacetic acids were separated in reversed-migration MEKCto the baseline. About 145-fold increases in detection sensitivity were observed with onlineconcentration strategy, compared with standard hydrodynamic injection (5 s at 25 mbarpressure). LODs ranged from 0.002 to 0.005 mg/L using only the online preconcentrationprocedures without any offline concentration of the extract. A DLLME procedure was usedin combination with the proposed online preconcentration strategies, which achieved thedetermination of analytes at limits of quantification ranging from 0.2 to 0.5 �g/kg, which isfar lower than the maximum residue limits established by China. The satisfactory recoveriesobtained by DLMME spiked at two levels ranged from 67.2 to 99.4% with RSD <15%, mak-ing this proposed method suitable for the determination of phenoxyacetic acids in watersamples.

Keywords: Dispersive liquid-liquid microextraction / Drinking water / Herbicides /Micellar electrokinetic chromatography / Phenoxyacetic acidsDOI 10.1002/jssc.201300290

1 Introduction

Phenoxyacetic acids, a conventional type of herbicide, havegained extensive applications in agriculture to control broad-leaved weeds, especially in developing countries includ-ing China [1]. Among these phenoxyacetic acids, 2,4-dichlorophenoxyacetic acid (2,4-D) is the most widely usedherbicide in the world (Industry Task Force Research Data).A number of studies had been carried out on the toxicity of2,4-D, and some researchers have demonstrated that 2,4-Dmight cause acute toxicity to common carp [2]. Therefore,

Correspondence: Professor Bining Jiao, Citrus Research InstituteSouthwest University/Laboratory of Citrus Quality and SafetyRisk Assessment, Ministry of Agriculture/Key Laboratory of Hor-ticulture Science for Southern Mountainous Regions of Ministryof Education, Southwest University, Chongqing 400715, ChinaE-mail: [email protected]: +86 23 68349046

Abbreviations: 2,4,5-T, 2,4,5-trichlorophenoxyaceticacid; 2,4-D, 2,4-dichlorophenoxyacetic acid; 2,6-D, 2,6-dichlorophenoxyacetic acid; 4-CPA, 4-chlorophenoxyaceticacid; DAD, diode array detection; DLLME, dispersiveliquid–liquid microextraction; LLE, liquid–liquid extraction;RM-MEKC, reversed-migration micellar electrokinetic chro-matography; SRMM, stacking with reverse migration ofmicelles

phenoxyacetic acids, an important pollution source in agri-cultural production, will cause potential pollution to soil andgroundwater with further potential toxicity against humans.Therefore, it is necessary to develop sensitive and selectivemethods for the analysis of phenoxyacetic acid residues inwater usually present at trace level. Analytical methods in-clude GC [3], GC–MS [4], HPLC [5, 6], HPLC–MS/MS [7],and CE [1, 8–10].

CE, the latest addition to analytical instrumentation, un-doubtedly plays an important role in the analysis of pesticideresidues due to its broad potential to separate a wide spec-trum of small molecules [11–13]. Among various CE modes,MEKC has prevailed as a powerful separation technique forboth neutral and ionic compounds [14]. The limitation of nor-mal or reverse MEKC in determining trace analytes is its lowsensitivity caused by the short path length in on-column UVdetection and by the low injection volumes (ca.10–50 nL). Dif-ferent on-capillary enrichment schemes with simplicity andeconomy have been developed to overcome this limitation.The most popular are sweeping (SW) and stacking with re-verse migration of micelles with (SRMM) or without (SRW)the insertion of a water plug prior to the sample and with orwithout polarity switching [13–15].

Quick and effective sample preparation techniques com-bined with the reliable analytical methods are necessary. Sam-ple preparation plays an important part in the research field of

C© 2013 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim www.jss-journal.com

3068 Y. Zhang and B. Jiao J. Sep. Sci. 2013, 36, 3067–3074

Figure 1. Structures of the four phenoxyacetic acids.

pesticide residue analysis and the commonly used methodsinclude liquid–liquid extraction (LLE) [16] and SPE [17, 18].For about a decade, a growing number of studies have fo-cused on two kinds of microextractions termed as liquid-phase microextraction (LPME) and solid-phase microextrac-tion (SPME), based on the miniaturization of conventionalLLE and SPE, respectively [19–23]. In 2006, Rezaee et al. de-veloped a new liquid–liquid microextraction method nameddispersive liquid–liquid microextraction (DLLME) [16]. Thenovel method has been widely recognized due to its sim-plicity, low cost, and high enrichment that made it availableto most analytical laboratories [24–28]. Unfortunately, therewere few reports on the extraction and enrichment of phe-noxyacetic acids in water using DLLME, while those conven-tional sample preparation techniques including LLE and SPEhave prevailed [4, 6–10].

In this paper, a new method was developed for the anal-ysis of four phenoxyacetic acids, 2,4,5-trichlorophenoxyaceticacid (2,4,5-T), 2,4-D, 2,6-dichlorophenoxyacetic acid (2,6-D),and 4-chlorophenoxyacetic acid (4-CPA), in drinking water.Figure 1 shows their structures. This method comprisedDLLME for offline preconcentration, SRMM for online pre-concentration, and CE combined with diode array detection(DAD) for qualification and quantification. The proposedmethod has a higher sensitivity than most of the recentlyprescribed methods with the commonly used detectors.

2 Materials and methods

2.1 Chemicals and standards

2,4-D (99.0% purity), 2,6-D (98.5% purity), 2,4,5-T (97.5%purity), and 4-CPA (99.0% purity) were from Dr. Ehrenstorfer.

HPLC-grade methanol and acetonitrile were fromCNW (Germany). Acetone, ethanol, chloroform, carbontetrachloride, tetrachloroethylene, 1,1,2-trichloroethane, andchlorobenzene were analytical reagents (Sinopharm Chem-cial Reagent, Shanghai, China). SDS (purity ≥ 98.5%) wasobtained from Sigma–Aldrich. Other reagents were of analyt-

ical grade. Phosphate buffer was prepared by mixing H3PO4

of appropriate concentration with 5 M NaOH to give therequired pH.

The stock solution of four phenoxyacetic acids was pre-pared at 1000 mg/L in methanol and stored in a glass vol-umetric flask at –50�C. It was completely stable for overthree months. Standard working solutions at a series of con-centrations were prepared by the dilution of aliquots of thestock solution.

All solutions were filtered through a membrane filter(0.22 �m) before use. Deionized water (18 M� cm resistiv-ity) from a Milli-Q Advantage A10 SP Reagent Water System(Millipore, Bedford, MA, USA) was used throughout.

2.2 Apparatus

A 7100 CE system (Agilent, USA) equipped with a DAD wasused. The separations were carried out in an uncoated fused-silica capillary (50 cm (40 cm to the detector) 75 mm id,Yongnian Optic Fiber, Hebei, China) at 25�C. An FE20 pHmeter (Mettler Toledo, Shanghai, China) was used for thepreparation of various buffers. A CL31R multispeed refrig-erated centrifuge (Thermo Scientific, USA) and GENIUS 3vortex agitator (IKA, Germany) were used for the samplepreparation.

2.3 DLLME procedure

Drinking water samples were collected and preserved ac-cording to the corresponding standard established by China(standard examination methods for drinking water, Regula-tion (GB/T) No. 5750–2006). Water samples were adjusted topH 2.0 by adding HCl. Five grams of the sample describedwas placed in a 15 mL conical centrifuge tube. The optimummixture of 1.00 mL of acetone (as dispersive solvent) and 90�L of 1,1,2-trichloroethane (as extraction solvent) was quicklyinjected into the tube and then emulsified to form a cloudysolution. Then the mixture was violently whirled for 5 minand centrifuged at 4000 rpm for 5 min. After that, the dis-persed fine droplets of 1,1,2-trichloroethane were sedimentedat the bottom of the tube. The sedimented phase was trans-ferred and concentrated under a stream of nitrogen at 45�Cto dryness. Finally, the residue was redissolved in 200 �L ofdeionized water and filtered through a 0.22 �m membranefilter for CE analysis.

2.4 CE procedure

New capillaries were successively pretreated with water,NaOH (1 M), water, HCl (1 M), and water, each for 30 min.Each day before starting analysis, the capillary was rinsedwith NaOH (0.1 M) and water for 10 min, and then precon-ditioned with running buffer for 10 min. Between runs, the


J. Sep. Sci. 2013, 36, 3067–3074 Sample Preparation 3069

capillary was flushed successively with NaOH (0.1 M), water,and running buffer, each for 5 min.

The CE running buffer consisted of a phosphate buffersolution of 20 mM at pH 2.1, containing 54 mM SDS and7.5% methanol. The sample was hydrodynamically injectedat 25 mbar for 5 s. A negative voltage (–20.0 kV) was applied,giving an average current of 68 �A.

For SRMM, the sample was diluted in plain water. Hydro-dynamic injections were performed at 50 mbar for 60–180 sat intervals of 10 s (optimum injection time 170 s). A positivevoltage was applied for 15–21 s at intervals of 1 s (optimuminjection time 19 s) at the inlet vial right (+20 kV). Then, thevoltage was switched to negative (–20 kV).

The stability of the reversed EOF was studied as follows.Acetone was used as a neutral EOF marker; 260 nm was se-lected as the detection wavelength. The solution of acetonewas prepared at 1% v/v in deionized water. The CE run-ning buffer was as mentioned above. The solution was hy-drodynamically injected at 25 mbar for 5 s and then positive(20.0 kV) or negative voltage (–20.0 kV) was applied. After240 min, the capillary was rinsed with the running buffer.The above experiment was repeated six times.

3 Results and discussion

3.1 Optimum separation conditions

The online preconcentration techniques tested in this paperrequired the separation of the analytes in reversed-migrationmicellar electrokinetic chromatography (RM-MEKC), inwhich the EOF direction is reversed to attain matrix removaland preconcentration. EOF can be reversed by two main path-ways: reduce pH to where EOF can be reduced but the elec-tophoretic mobility of the anionic micelles will not be over-whelmed or add a cationic surfactant that is adsorbed in thepositively charged capillary walls [14, 15, 29].

The separation of four phenoxyacetic acids in RM-MEKCwas optimized by testing different running buffers mixedwith phosphate solutions (15–40 mM) at different pH (1.9–2.4), SDS concentrations (24–60 mM), and different solvents(methanol, acetonitrile, isopropanol) between 5 and 15% v/vby direct hydrodynamic injection (5 s at 25 mbar) of a standardsolution at 5.0 mg/L prepared in the running buffer. Amongthe tested running buffers, that containing 20 mM phosphate(pH 2.1), 54 mM SDS, and 7.5% methanol v/v provided thebest results.

Finally, the effect of the applied voltage (15.0–25.0 kV)and temperature (20–30�C) was studied to shorten analysistime as much as possible. A voltage of 20.0 kV at 25�C was theoptimum result. Figure 2 shows the electropherogram of thestandard solution (5.0 mg/L) under the optimized conditions.Four phenoxyacetic acids are separated to the baseline within6 min. In addition, the maximum adsorption of 2,4,5-T,2,4-D, 2,6-D, and 4-CPA were at about 206, 202, 198, and194 nm, respectively. Regarding the sensitivities of the ana-lytes, 200 nm was selected as the detection wavelength.

In addition, the stability of the reversed EOF was studied.Acetone was used as a neutral EOF marker. The result showedthat there was no peak under positive or negative voltage in240 min and then an obvious peak under pressure. Run-to-run reproducibility was excellent. An earlier study suggestedthat a low EOF corresponding to an elution time >120 min wasassigned as a zero flow rate [30]. So, the EOF was consideredas zero flow rate in our experiment.

3.2 Online sample preconcentration strategies

In our experiment, SRMM was used for the online precon-centration strategy. In a systematic investigation, it has beenfound that hydrodynamic injection time and dynamoelectricinjection time play an important role in SRMM online precon-centration. Figure 3A illustrates the effect of hydrodynamicinjection time on peak area for SRMM. The peak areas ofthe analytes increased in proportion to the injection time.However, wider peaks and longer migration times appearedbeyond 170 s, although the peak areas of most analytes hadan uptrend. Figure 3B illustrates the effect of the dynamo-electric injection time on peak area for SRMM. At first, thepeak areas increased gradually with the increase of the dy-namoelectric injection time. The maximal peak areas of fourphenoxyacetic acids were all obtained at 19 s. And then theydropped down slowly with the increase of the dynamoelectricinjection time. So, 170 and 19 s were chosen as the optimumhydrodynamic injection time and dynamoelectric injectiontime, respectively.

Figure 4 shows the electropherograms obtained bySRMM–RM-MEKC–DAD analysis of a standard mixtureat 1 mg/L. A usual injection is included for comparison.About 149-, 148-, 143-, and 144-fold increases in detection

Figure 2. Electropherogram of a standard mixture at 5 mg/Lby RM-MEKC in 20 mM phosphate buffer at pH 2.1 containing54 mM SDS and 7.5% methanol v/v. Other conditions: hydrody-namic injection of 5 s at 25 mbar pressure; voltage of –20 kVat 25�C; detection wavelength of 200 nm. Peak identification:1 = 2,4,5-T; 2 = 2,4-D; 3 = 2,6-D; 4 = 4-CPA.



Figure 3. Effects (A) of hydrodynamic injection time and (B) ofdynamoelectric injection time on the peak areas of analytes forSRMM: (A) the dynamoelectric injection time was 16 s; (B) thehydrodynamic injection time was 170 s.

sensitivity, compared with the standard hydrodynamic injec-tion, were observed for SRMM for the four acids.

Table 1 summarizes and compares the mean retentiontimes, linearity, LODs, RSD% of peak area, and migrationtimes for normal injection 5 s at 25 mbar pressure andSRMM, respectively. The linearity of these methods wastested using five different concentrations within the rangeof 0.002–2 mg/L, executing at least three replication injec-tions. The results reveal a satisfactory linearity for all theanalytes with the correlation coefficients (R2) higher than0.990 in the linear regression equation. Data on the regres-sion equations are listed in Table 1. The LODs (S/N = 3)of the four compounds were between 0.002 and 0.005 mg/L(SRMM). The reproducibility (RSDs) of the proposed meth-ods in terms of peak area for six replicate injections, were4.9–8.4% for SRMM. The RSDs in terms of migration timewere between 2.4 and 3.7% for SRMM.

Figure 4. Electropherograms of a standard mixture at 1 mg/L (A)by conventional RM-MEKC with the hydrodynamic injection (5 s at25 mbar) and (B) by the SRMM–RM-MEKC. SRMM: hydrodynamicinjection 170 s at 50 mbar, application of positive voltage 19 s.Separation, other conditions, and peak identification as in Fig. 2.

3.3 Optimization of DLLME

When applying this method to drinking water, sample pre-concentration by DLLME was required (with about 25-foldpreconcentration achieved). As is well known, the properextractant and dispersant are the top-drawer factors in theDLLME process. Therefore, the type and volume of extrac-tion solvent or dispersive solvent should be first optimized.

3.3.1 Selection of extraction solvent

Halogenated organic compounds are representative extrac-tion solvents [31]. Dispersive solvents should be miscible inboth aqueous samples and extraction solvents to help theanalytes to transfer successfully from the former phase intothe latter. Thus, acetone, acetonitrile, methanol, and ethanolare typical dispersive solvents. In our study, five commonhalogenated solvents, chloroform (CHCl3), carbon tetrachlo-ride (CCl4), tetrachloroethylene (C2Cl4), 1,1,2-trichloroethane



Table 1. LODs, RSDs (n = 6) and linearity (five points, injections in triplicate) for the studied phenoxyacetic acids with RM-MEKC–DADusing conventional injection and SRMM

Phenoxyacetic acids

2,4,5-T 2,4-D 2,6-D 4-CPA

Conventional injectionMean migration time (min) 4.206 4.575 5.217 5.505Migration time (RSDs, %) 2.5 1.8 2.6 3.4Peak area (RSDs, %) 6.8 7.1 5.5 6.3Linear range (mg/L) 0.4–10 0.2–10 0.2–10 0.2–10Linear equation y = 0.96x – 0.17 y = 1.23x – 0.14 y = 1.56x – 0.19 y = 1.46x – 0.24Correlation coefficient (R2) 0.994 0.995 0.993 0.992LODs (mg/L) 0.4 0.2 0.2 0.2

SRMMMean migration time (min) 6.878 7.217 7.824 8.049Migration time (RSDs, %) 3.7 2.4 3.1 2.9Peak area (RSDs, %) 4.9 8.4 6.7 7.4Linear range (mg/L) 0.005–2 0.002–2 0.002–2 0.002–2Linear equation y = 109.80x + 4.14 y = 144.66x + 4.82 y = 171.57x + 5.97 y = 159.56x + 5.95Correlation coefficient (R2) 0.991 0.992 0.993 0.992LODs (mg/L) 0.005 0.002 0.002 0.002

(C2H3Cl3), and chlorobenzene (C6H5Cl), were selected as ex-tractants.

A series of extraction solvents were evaluated for enrich-ment as follows: 5.00 g of drinking water (pH 2.0) was spikedwith the standard solution at 0.01 mg/kg, 1.00 mL of acetonewas used as dispersive solvent and the volumes of extractionsolvents were 50 �L. The Extraction efficiency was evaluatedby comparing the peak areas of the analytes. The effects offive extraction solvents on the peak areas of the analytes areshown in Fig. 5A. The results show that 1,1,2-trichloroethaneproduced a higher extraction recovery than the other four ex-traction solvents. Thus, 1,1,2-trichloroethane was selected asthe optimum extraction solvent.

3.3.2 Effect of extraction solvent volume

In order to investigate the volume effect of the extraction sol-vent, C2H3Cl3 was changed from 50 to 100 �L in the intervalof 10 �L, while the dispersive solvent (acetone) was main-tained at 1.00 mL. The volume effect of C2H3Cl3 on the peakareas of the analytes is displayed in Fig. 5B. The extractionefficiency was improved with the increase of volume. WhenC2H3Cl3 was increased from 50 to 90 �L, the peak areas ofthe analytes were gradually enhanced. The maximal peak ar-eas were all obtained at 90 �L. As a result, 90 �L of C2H3Cl3was selected as the optimum extraction volume to provide asatisfactory recovery.

3.3.3 Selection of dispersive solvent

Acetone, acetonitrile, methanol, and ethanol were chosen asthe dispersive solvents. A series of dispersive solvents wereevaluated as follows: 5.00 g of drinking water (pH 2.0) wasspiked with the standard solution at 0.01 mg/kg, 90 �L of

C2H3Cl3 was used as extraction solvent, and the volumes ofdispersive solvents were 1.00 mL. The extraction efficiencywas estimated by comparing the peak areas of the analytes.The effects of dispersive solvents on the peak areas of theanalytes are shown in Fig. 6A. The results show that acetoneproduced higher extraction recovery than the other three dis-persive solvents. Therefore, acetone was chosen as the opti-mum dispersive solvent.

3.3.4 Effect of dispersive solvent volume

In order to study the volume effect of the dispersive solvent,acetone was varied from 0.6 to 1.6 mL in the interval of 0.2 mL,while the extraction solvent (C2H3Cl3) was kept at 90 �L. Asshown in Fig. 6B, with the increase of acetone from 0.6 to1.0 mL, the extraction efficiency increased gradually anddropped down slightly >1.0 mL. Therefore, 1.0 mL of ace-tone was selected as the optimum dispersive solvent volumeto obtain an acceptable recovery.

3.3.5 Effect of pH

In general, analytes in neutral forms are much easier to ex-tract than those in ionic forms [32]. Consequently, analytesare often expected to be in a nonionic state in DLLME forhigher extraction efficiency. Because pH can affect the formsof compounds in solution, the effect of the sample pH on ex-traction efficiency was investigated. The pH was varied from1.8 to 3.4 in the interval of 0.2 and the other conditions ofDLLME were as described. The effects of pH on the peakareas of the analytes are shown in Fig. 7. At first, with theincrease of pH, the peak areas of all the analytes increasedslightly. After the peak areas reached their highest, they de-creased obviously; pH 2.0 was chosen to obtain the maximum



Figure 5. Effects (A) of extraction solvents and (B) of the volumesof extraction solvents on peak areas of analytes in DLLME; 5.00 gof spiked drinking water at 0.01 mg/kg.

recoveries of most of the analytes. This can be explained bythe pKa values of 2,4,5-T, 2,4-D, 2,6-D, and 4-CPA of 2.85,2.73, 2.95 and 3.56, respectively.

3.3.6 Effect of vortex time and centrifugation time

Centrifugation time has no obvious influence on the extrac-tion efficiency in DLLME [33]. To study their effects, the vortextime and centrifugation time were changed at a 1 min inter-val from 3 to 7 min, while the centrifugation was fixed at4000 rpm. The other conditions of DLLME were describedabove. The results revealed that both vortex time and cen-trifugation time had a slight effect on the extraction efficiency.Thus, a vortex time of 5 min and centrifugation time of 5 minwere chosen to help the cloudy solution to precipitate at thebottom of the tube.

Figure 6. Effects (A) of dispersive solvents and (B) of the vol-umes of dispersive solvents on peak areas of analytes in DLLME;5.00 g of spiked drinking water at 0.01 mg/kg.

Figure 7. Effect of pH on peak areas of analytes in DLLME.



Table 2. Accuracy, precision, and LOQs obtained after DLLME andonline preconcentration of spiked drinking water usingSRMM-RM-MEKC

Analytes Added(mg/kg)

Recovery, %(RSD, n = 6)

LOQs(�g/kg)

Concentrationratio (only byDLLME)

2,4,5-T 0.002 96.8 (12.5) 0.4 240.01 99.4 (9.1) 25

2,4-D 0.002 84.2 (13.1) 0.2 210.01 88.7 (10.8) 22

2,6-D 0.002 74.4 (14.6) 0.5 190.01 77.6 (11.7) 19

4-CPA 0.002 67.2 (12.3) 0.5 170.01 68.1 (10.4) 17

3.4 Method validation

Table 2 shows the accuracy of this method expressed as themean recoveries and the precision expressed as the corre-sponding RSDs. The recoveries achieved from spiking wa-ter at two concentration levels of each phenoxyacetic acid(0.002 and 0.01 mg/kg) ranged from 67.2 to 96.8% (RSDs12.3–14.6%) and from 68.1 to 99.4% (RSDs 9.1–11.7%), re-spectively. The recoveries were slightly <70%, especially for4-CPA. Hence we can conclude that this method providessatisfactory accuracy and precision. In addition, the LOQ

Figure 8. Electropherogram of spiked drinking water at0.01 mg/kg after DLLME. Peak identification: 1 = 2,4,5-T; 2 = 2,4-D;3 = 2,6-D; 4 = 4-CPA.

(S/N = 10) of 2,4-D was undoubtedly far below the maxi-mum residue limit (MRL) of 0.03 mg/L established by China(standards for drinking water quality, Regulation (GB/T) No.5749–2006).

Figure 8 shows the chromatogram of the spiked water at0.01 mg/kg of each phenoxyacetic acid obtained after DLLME.Obviously, DLLME is a very simple and effective methodfor preconcentrating phenoxyacetic acid residues in drinkingwater. Additionally, there was no peak interference in thetypical chromatogram of blank water after DLLME.

Table 3. Comparison of the proposed method and some other methods for the determination of phenoxyacetic acids

CE mode Detection Online precon-centrationstrategies

Samplepreparation

Analyte Sample LOD or LOQ Recovery Ref.

CZE MS – – MCPP, 2,4-DP,MCPA, 2,4-D

– 20 pg – [34]

CZE combinedwith precolumnderivatization

LIF – – 2,4-D, 2,4,5-T,2-PPA,mecoprop,2,2-CPPA,2,3-CPPA,2,4-CPPA,dichloprop, silvex

– 0.2 �g/L – [35]

MEKC combinedwith precolumnderivatization

LIF – LLE Dicamba, 2,4-D,chlorimuronethyl

Pond water 10 ng/L – [9]

MEKC UV (200 nm) – – 2,4-D, 4-chloro-2-methylphen-oxyaceticacid

– 140–150 �g/L – [36]

MEKC LIF – SPE 2,4-D, 2,4-DB,2,4-DP, 2,4,5-TP,MCPA, MCPB

Water 2 fg – [10]

MEKC UV (210 nm) – SPE diuron,monolinuron,2,4-DP, 2,4,5-T

Surfacewater

2–5 �g/L 92–98% [8]

RM-MEKC DAD (200 nm) SRMM DLLME 2,4-D, 2,6-D, 2,4,5-T,4-CPA

Drinkingwater

0.2–0.5 �g/kg 67–99% Proposedmethod



Table 3 summarizes the details of the proposed methodand the other methods that were applied for phenoxyaceticacid determination. Clearly, this method has higher sensi-tivity than most of the recently prescribed methods with thecommonly used detectors, such as UV and DAD [8–10,34–36].

3.5 Application to real samples

The proposed method has been applied to the analysis ofreal samples. However, no phenoxyacetic acid was detectedin dozens of drinking water samples. The future introductionof other kinds of water samples, such as lake water and riverwater, may help to find more widespread application.

4 Concluding remarks

A new method using RM-MEKC–DAD was developed to de-tect four phenoxyacetic acids. The online preconcentrationstrategy was used in combination with DLLME as offline pre-concentration. The proposed method has been successfullydemonstrated in the analysis of phenoxyacetic acids in drink-ing water with good recoveries from 67 to 99%.

This work was supported financially by Special Fund of Tech-nological Innovation System Construction of Modern Agriculture(CARS-27) and Fundamental Research Fund for the CentralUniversities (NO XDJK2012C059).

The authors have declared no conflict of interest.

5 References

[1] Fu, F. F., Xiao, L. X., Wang, W., Xu, X. Q., Xu, L. J.,Qi, G. M., Chen, G. N., Sci. Total Environ. 2009, 407,1998–2003.

[2] Sarikaya, R., Yılmaz, M., Chemosphere 2003, 52,195–201.

[3] Erickson, L. C., Hield, H. Z., J. Agric. Food Chem. 1962,10, 204–207.

[4] Kuang, K., Zhao, J. L., Ying, G. G., Yuan, Y. Q., Yang, X. B.,Li, Z. J., J. Instrumental Anal. 2008, 27, 816–819.

[5] Shamsipur, M., Fattahi, N., Pirsaheb, M., J. Sep. Sci.2012, 35, 2718–2724.

[6] Tran, A. T. K., Hyne, R. V., Doble, P., Chemosphere 2007,67, 944–953.

[7] Pozo, O., Pitarch, E., Sancho, J. V., J. Chromatogr. A 2001,923, 75–85.

[8] Farrn, A., Ruiz, S., J. Chromatogr. A 2004, 1024, 267–274.

[9] Liu, H. X., Song, J. M., Han, P., J. Sep. Sci. 2006, 29,1038–1044

[10] Wall, W., Li, J., El Rassi, Z., J. Sep. Sci. 2002, 25,1231–1244.

[11] Karcher, A., El Rassi, Z., Electrophoresis 1999, 20,3280–3296.

[12] Malik, A. K., Faubel, W., Crit. Rev. Anal. Chem. 2001, 31,223–279.

[13] Hernandez-Borges, J., Frıas-Garcıa, S., Cifuentes, A.,Rodrıguez-Delgado, M. A., J. Sep. Sci. 2004, 27,947–963.

[14] Shihabi, Z. K., J. Chromatogr. A 2000, 902, 107–117.

[15] Quirino, J. P., Kim, J. B., Terabe, S., J. Chromatogr. A2002, 965, 357–373.

[16] Rezaee, M., Assadi, Y., Hosseini, M. R. M., Aghaee, E.,Ahmadi, F., Berijani, S., J. Chromatogr. A 2006, 1116,1–9.

[17] Hernandez, F., Sancho, J. V., Pozo, O., Lara, A., Pitarch,E., J. Chromatogr. A 2001, 939, 1–11.

[18] Sharif, Z., Man, Y. B. C., Hamid, N. S. A., Keat, C. C., J.Chromatogr. A 2006, 1127, 254–261.

[19] Paillakis, E., Kalogerakis, N., Trends Anal. Chem. 2003,22, 565–574.

[20] Paillakis, E., Kalogerakis, N., Trends Anal. Chem. 2002,21, 53–63.

[21] Rasmussen, K. E., Pedersen-Bjergaard, S., Trends Anal.Chem. 2004, 23, 1–10.

[22] Correia, M., Delerue-Matos, C., Alves, A., Fresenius J.Anal. Chem. 2001, 369, 647–651.

[23] Zambonin, C. G., Cilenti, A., Palmisano, F., J. Chromatogr.A 2002, 967, 255–260.

[24] Tolcha, T., Merdassa, Y., Megersa, N., J. Sep. Sci. 2013,36, 1119–1127.

[25] Mokhtari, B., Dalali, N., Pourabdollah, K., J. Sep. Sci.2013, 36, 356–361.

[26] Montes, R., Rodriguez, I., Ramil, M., Rubi, E., Cela, R., J.Chromatogr. A 2009, 1216, 5459–5466.

[27] Ravelo-Perez, L. M., Hernandez-Borges, J., Asensio-Ramos, M., Rodrıguez-Delgado, M. A., J. Chromatogr.A 2009, 1216, 7336–7341.

[28] Abdolmohammad-Zadeh, H., Sadeghi, G. H., Anal. Chim.Acta 2009, 649, 211–217.

[29] Kim, J. B., Quirino, J. P., Otsuka, K., Terabe, S., J. Chro-matogr. A 2001, 916, 123–130.

[30] Sui, Z. J., Schlenoff, J. B., Langmuir 2003, 19,7829–7831.

[31] Rezaee, M., Yamini, Y., Faraji, M., J. Chromatogr. A 2010,1217, 2342–2357.

[32] Lu, Q. M., Chen, L. H., Lu, M. H., Chen, G. N., Zhang, L.,J. Agric. Food Chem. 2010, 58, 2763–2770.

[33] Boonchiangma, S., Ngeontae, W., Srijaranai, S., Talanta2012, 88, 209–215.

[34] Nielen, M. W. F., J. Chromatogr. A 1995, 712,269–284.

[35] Mechref, Y., El Rassi, Z., Anal. Chem. 1996, 68,1771–1777.

[36] Farran, A., Serra, C., Sepaniak, M. J., J. Chromatogr. A1999, 835, 209–215.


Dispersive liquid–liquid microextraction combined with high-performance liquid chromatography–tandem mass spectrometry for the identification and the accurate quantification by

Documents