Analytica Chimica Acta 433 (2001) 207–215
Solventless sample preparation procedure for organophosphoruspesticides analysis using solid phase microextraction and on-line
supercritical fluid extraction/high performance liquidchromatography technique
Shamsul Hairi Salleha, Yoshihiro Saitoa, Yoshiaki Kisob, Kiyokatsu Jinnoa,∗a School of Materials Science, Toyohashi University of Technology, Tempaku-cho, Toyohashi 441-8580, Japan
b School of Ecological Engineering, Toyohashi University of Technology, Tempaku-cho, Toyohashi 441-8580, Japan
Received 10 October 2000; received in revised form 18 December 2000; accepted 05 January 2001
Abstract
Organophosphorus pesticides spiked in water sample solutions were extracted with solid phase microextraction (SPME)technique and were desorbed by supercritical fluid carbon dioxide (SFCO2) before on-line introduction into high performanceliquid chromatography system (HPLC). All of the 5-ml SFCO2 injected into the HPLC system were dissolved into themethanol/water (80/20) mobile phase without affecting the baseline of the UV detector signal. Preliminary experiments wereperformed in order to optimize the extraction conditions of pesticides from spiked water solutions with the SPME technique.Subsequently, desorption conditions of the pesticides from the SPME fiber coating using SFCO2 as the desorption mediumwere optimized. © 2001 Elsevier Science B.V. All rights reserved.
Keywords:Organophosphorus pesticides; Solid phase microextraction; Supercritical fluid extraction; Solventless sample preparationtechnique; On-line coupling
1. Introduction
Sample preparation procedures prior to the chro-matographic analyses is one of the most critical stepsin an analytical process. Nevertheless, it can also be themost time and labor intensive, usually has multi-stepprocedures prone to loss of analytes, and uses toxicorganic solvents. Consequently, most of the analysistime is spent on the sample preparation steps alone.Various extraction techniques for sample preparationin chromatographic analyses have been developedin order to overcome these problems. Among these
∗ Corresponding author. Fax:+81-532-48-5833.E-mail address:[emailprotected] (K. Jinno).
techniques, solid phase microextraction (SPME)and supercritical fluid extraction (SFE) seem verypromising.
Extensive studies have shown SPME to be an effec-tive technique for sample preparation procedure. Thistechnique has been successfully employed to analyzea wide range of pollutants, such as polycyclic aromatichydrocarbons (PAHs) [1], phenols [2,3], BTEX (ben-zene, toluene, ethylbenzene and xylenes) [4,5] andalso pesticides [6,7]. SPME was also coupled to highperformance liquid chromatography (HPLC) by a des-orption chamber and a six port valve interface in orderto apply this sample preparation technique to a widerrange of compounds, i.e. non-volatile or thermally un-stable compounds [8] which are normally unsuitable to
0003-2670/01/$ – see front matter © 2001 Elsevier Science B.V. All rights reserved.PII: S0003-2670(01)00790-5
208 S.H. Salleh et al. / Analytica Chimica Acta 433 (2001) 207–215
gas chromatography (GC) analysis. The SPME/HPLCtechnique was also automated (in-tube SPME)[9,10] and coupled to electrospray/mass spectrometry(ESI/MS) [11] to achieve higher sensitivity and lowerdetection limits. More recent work on the in-tubeSPME technique describing the use of tailor-madefused silica capillaries coated on the inside withpolysilicone acrylate was also reported by Hartmannet al. [12]. Other new approaches on SPME/HPLCwere also reported by the authors' group [13,14].
On the other hand, SFE that has been utilized pri-marily in large-scale industrial applications is beingpaid much attention as an analytical sample prepara-tion technique in the past 10 years. This technique notonly offers advantages such as better extraction effi-ciencies, reduced time of extraction and most of all,reduced amount of organic solvents but also offers adiverse types of compounds that could be used as thesupercritical fluid (SF). The on-line coupling of SFEto various types of chromatographic techniques havebeen reported by many researchers. Most of them fo-cused on the on-line coupling between SFE and GCand between SFE and supercritical fluid chromatog-raphy (SFC) [15–17]. Lanças and Ruggiero [18] alsodescribed on-line coupling of SFE to capillary columnelectrodriven separation techniques. However, only afew reports have been published describing on-linecoupling of SFE and HPLC [19–21] and most of thetechniques employed an impactor interface for trap-ping the desorbed analytes. Consequently, organic sol-vents have to be used to elute the analytes trapped onthe impactor.
Ashraf-Khorassani et al. [22] reported an on-linecoupling of SFE with HPLC for the analysis of fourPAHs and five linear alkylbenzene sulfonates. In theirstudy, the supercritical fluid containing extracted ana-lytes was injected directly into the HPLC system with-out any impactor (packed bed) interface for extractedanalytes collection. They demonstrated that all of the9-ml supercritical carbon dioxide (SFCO2) injectedinto the HPLC system were dissolved in the mobilephase when methanol/water mobile phases from 100to 80% were used and did not adversely affect thebaseline of the UV detector signal.
In accordance with the objective of our previousstudies [23,24], i.e. the development of a solventlesssample preparation method, and to explore the advan-tages of both techniques, SPME and SFE, we have
constructed a desorption system which enables SFCO2containing desorbed pesticides from the SPME fibercoating to be injected directly into the HPLC system.Consequently, no collection solvent or impactor in-terface is required for the trapping of the pesticidesdesorbed by the SFCO2. Furthermore, this techniquewould be found useful when dealing with extensivelycontaminated samples, since selective desorption ofthe target analytes could be performed by adjustingthe desorption pressure and/or temperature of the su-percritical fluid which could not be done with the nor-mal organic solvent desorption process. Consequently,co-extractant from the samples would be left on theSPME fiber coating and cleaner chromatograms couldbe obtained. Swelling of the polymeric coating of theSPME fiber due to organic solvent frequently led tothe stripping of the coating from the fused-silica fiberwhen the SPME assembly is withdrawn through thedouble tempered ferrule. This problem could also beavoided with this technique since no organic solventis used in desorption process subsequently providinglonger lifetime of the SPME fiber coating.
In this study, the SPME of the pesticides was per-formed off-line and after the end of the extraction pe-riod, the SPME manual assembly was installed ontothe desorption device for the desorption process. TheSFCO2 containing desorbed pesticides was then intro-duced into the HPLC system on-line.
2. Experimental
2.1. Instrumentation
The on-line introduction of the desorbed pesticidesfrom the SPME coating into the chromatographicsystem was accomplished by employing the des-orption system depicted in Fig. 1(a) and (b). Thesystem consisted of Supelco Supercritical Fluid Ex-tractor SFE-400 (Bellefonte, PA, USA), two Rheo-dyne 7125 injectors (Cotati, CA, USA), Luna-ODS(150 mm × 4.6 mm i.d., particle size 3mm, Phe-nomenex, Torrance, CA, USA) analytical column,Jasco PU-980 pump and Jasco MD-915 photodiodearray detector (Jasco, Tokyo, Japan). The conditionsfor the analysis of pesticides were as follows: — col-umn temperature: 35◦C; mobile phase: a mixture ofmethanol and water (80/20); flow rate: 1.0 ml min−1.All data were electronically acquired and processed
S.H. Salleh et al. / Analytica Chimica Acta 433 (2001) 207–215 209
Fig. 1. (a) The inside view of the oven (D) showing the SPME manual assembly installed on the desorption device. (b) Schematic diagramof on-line SFE/HPLC — (A): CO2 gas cylinder; (B): two-way valve B; (C): supercritical fluid extractor; (D): oven; (E): two-way valveE; (F): injection valve F; (G): injection valve G; (H): 5-ml injection loop; (I): 20-ml injection loop; (J): 100-ml stainless steel tube; (K):two-way valve K.
using BORWIN chromatography software (Jasco,Tokyo, Japan) on a personal computer. The super-critical fluid extractor houses a thermal pump thatheats the liquid carbon dioxide to achieve the elevatedpressure. Therefore, no conventional pump was usedfor the delivery of the SFCO2 into this system. Thedesorption chamber (Supelco, Bellefonte, PA, USA),as shown in Fig. 1(a), consists of three parts: a stain-less steel body (internal volume of 70ml), a stainlesssteel cap and double tempered ferrule which is placedbetween them. The original clamp used to compressthe stainless steel cap was removed and two U-shapestainless steel bars were employed to withstand highpressure.
The hub of the SPME fiber assembly was foundnot strong enough to withstand the high pressure thatgave cracks and leakage of the SFCO2 from the des-orption chamber. Since the stainless steel microtubingthat holds the fiber coating is an open tubular stainlesssteel tube, the pressurized CO2 could enter through themicrotubing and causes cracks at the hub. Therefore,
modification of the SPME fiber assembly was made asshown in Fig. 2. The stainless steel microtubing wasdetached from the hub and the end of the microtubingwas soldered to serve as a seal for the microtubing.The microtubing was then reattached to the hub withepoxy glue.
2.2. Procedure
To perform the extraction of pesticides from spikedwater solutions with SPME technique, 85mm of poly-acrylate coating (Supelco, Bellefonte, PA, USA) waschosen as the medium. The SPME fiber was immersedin the 15-ml spiked water solutions for a period oftime. The spiked water solutions were stirred with8 mm× 10 mm o.d. star shape stirrer with a magneticstirrer (Iuchi, Osaka, Japan). During all extractions,spiked water solutions were kept to the set tempera-ture by immersing the sample vial in a water bath.
After the completion of the extraction of the pesti-cides with the SPME technique, the fiber coating was
210 S.H. Salleh et al. / Analytica Chimica Acta 433 (2001) 207–215
Fig. 2. Modification made on the SPME fiber assembly.
let to dry by exposing it to the air for 15 min. This is toremove any moisture remained on the fiber coating thatcould effect the precision of this method. The SPMEmanual assembly with the fiber withdrawn was theninstalled on the desorption device in the supercriticalfluid extractor oven (D) as shown in Fig. 1(a). After theinstallation, the plunger of the SPME assembly waspushed downward, exposing the fiber coating beforetightening the U-shape stainless steel bars. There wasa period of 10 min before filling the desorption devicewith SFCO2 in order to equilibrate to the set tempera-ture in the oven. The desorption device was filled withSFCO2 by opening valve B and closing valve E. Open-ing valve E after a period of time delivers the SFCO2containing desorbed pesticides into the 5-ml injectionloop (H). At that time, both injection valves F and Gwere set to the load position. By turning valve F andG simultaneously to the inject position, will wash theSFCO2 containing desorbed pesticides into the HPLCsystem. By opening the two-way valve K, the SFCO2remained in the stainless steel tubing between G (6)and K will be released. The U-shape stainless steelbars were then removed, the fiber was drawn back intothe syringe needle, and the SPME manual assemblywas removed from the desorption device.
2.3. Reagents
Sodium chloride and solvents were reagent gradepurchased from Kishida Chemical (Osaka, Japan).Deionized water obtained from Milli-Q water sys-tem (Millipore, Tokyo, Japan) was used throughoutthe study. Four organophosphorus pesticides (ben-sulide, diazinon, EPN and chlorpyrifos) selected forthis study were obtained from Wako Pure Chemicals(Osaka, Japan). Their methanol stock solutions wereprepared at the concentration of 1000 mg l−1. Spikingstandards were prepared by mixing and diluting thissolution in methanol. For the extraction of pesticideswith SPME technique, 15-ml spiked water solutionscontaining 1 mg l−1 of each pesticide in 16.4-mlscrew cap vial with hole cap and Teflon faced septawere used, and the methanol content was kept below1% to avoid interference in the SPME process.
3. Results and discussion
3.1. Optimization of extraction conditions
The optimization of the SPME conditions of thepesticides has been carried out in our previous study[24] and the results showed that the pesticides tooka considerable time to reach equilibrium, i.e. morethan 180 min, which is generally impractical for realanalysis. Owing to this fact, we attempted prelimi-nary experiments to re-optimize the extraction condi-tions. In this study, a star shape stirrer was used in-stead of the normal shape stirrer bar in order to achievesufficient stirring of the spiked water solution. Theparameters tested for the optimization of the SPMEof the pesticides were temperature: 40–70◦C; extrac-tion time: 30–240 min; sodium chloride concentration:0.0–0.2 g ml−1 and the optimum conditions obtainedfrom these experiments are shown in Table 1. Fig. 3shows the peak area versus extraction time performedon the pesticides with the optimized conditions. As ex-pected, employing the star shape stir bar gave shortertime for the analytes to equilibrate. Although the equi-librium is not reached at 60 min for EPN, the extrac-tion time was set at 60 min for the optimization of thedesorption conditions for practical convenience. Theoptimized extraction conditions listed in Table 1 werethen used in all experiments to optimize the desorp-tion conditions.
S.H. Salleh et al. / Analytica Chimica Acta 433 (2001) 207–215 211
Table 1Optimized extraction and desorption conditions of pesticides established in this study
Extraction Desorption
Parameters Conditions Parameters Conditions
Temperature (◦C) 75 Pressure (atm) 306Time (min) 60 Temperature (◦C) 50NaCl concentration (g ml−1) 0.0 Static (min) 30
SFCO2 injection volume (ml) 5
3.2. Optimization of desorption conditions
It has been reported that SFCO2 is miscible inorganic solvents while SFCO2 has the solubility ofonly 6% in water [25]. Furthermore, as mentionedin Section 1, Ashraf-Khorassani et al. [22] demon-strated that by using methanol/water mobile phasesfrom 100 to 80% methanol in water with the flowrate of 1.0 ml min−1, all of the 9-ml SFCO2 injectedinto the HPLC system were dissolved in the mobilephase and no negative peak appeared in the baseline.However, as the percentage of the water in the mo-bile phase increased (more than 20%), the SFCO2solubility in the mobile phase decreased while theundissolved portion eluted like a gas which appeared
Fig. 3. Effect of extraction time on peak area of pesticides desorbedfrom SPME fiber coating. Concentration of pesticides: 1.0 mg l−1.Extraction conditions — temperature: 75◦C. Desorption conditions— pressure: 306 atm; temperature: 50◦C; static desorption: 30 minand SFCO2 injection volume: 5ml.
as a negative peak in a baseline of UV detectorsignal. In virtue to this fact, we firstly examinedthe effect of the mobile phase composition on thepeak shape of the chromatogram of 1 mg l−1 EPN.Two mobile phase compositions of acetonitrile/water(60/40) and methanol/water (80/20) were tested. Theresult is depicted in Fig. 4. From the chromatogram,it is evident that the mobile phase composition ofmethanol/water (80/20) as D in Fig. 4 gave sharperpeak shape compared to acetonitrile/water (60/40)as C in Fig. 4. Preliminary experiments also showed
Fig. 4. Chromatograms showing the effect of the mobile phasecomposition and SFCO2 injection volume on the peak shapeof 1.0 mg l−1 EPN. (A) Methanol/water (80/20) with 10-ml in-jection loop; (B) acetonitrile/water (60/40) with 10-ml injectionloop; (C) acetonitrile/water (60/40) with 5-ml injection loop; (D)methanol/water (80/20) with 5-ml injection loop. Extraction con-ditions — time: 60 min and temperature: 75◦C. Desorption condi-tions — pressure: 306 atm; temperature: 50◦C and static desorp-tion: 30 min.
212 S.H. Salleh et al. / Analytica Chimica Acta 433 (2001) 207–215
that peak broadening could not be resolved even withacetonitrile/water (80/20) suggesting that acetonitrileand methanol have different solvent strength whenthey are in the same composition. Consequently,acetonitrile is considered unsuitable for dissolvingSFCO2 compared to methanol and therefore the mo-bile phase composition of methanol/water (80/20)was used for all subsequent experiments. The effectof the injection loop volume on the peak shape wasalso investigated by changing the injection volumeto 5 and 10ml. Similarly, two mobile phase compo-sitions mentioned above were tested. Although the10-ml loop gave larger peak compared to 5-ml withacetonitrile/water (60/40) as B in Fig. 4, it also gave abroader peak. As for methanol/water (80/20) mobilephase composition, CO2 gas bubbles could be seeneluting out of the SFE/HPLC system, which gave asmaller and broader peak, as A in Fig. 4, indicatinginsufficient solvation of SFCO2 in the mobile phase.Consequently, the injection volume of 5ml was usedfor all subsequent experiments.
Fig. 5. Chromatograms showing the non-hom*ogenousity of thedesorbed pesticides in the desorption chamber. Volume of thestainless steel tube from the desorption chamber to the valve E(A) 69ml and (B) 42ml. Concentration of pesticides: 1.0 mg l−1.Extraction conditions — time: 60 min and temperature: 75◦C.Desorption conditions — pressure: 306 atm; temperature: 50◦C;static desorption: 30 min and SFCO2 injection volume: 5ml.
Preliminary experiments also showed that the des-orbed analytes from the SPME fiber coating did notdisperse hom*ogeneously in the desorption device. Bychanging the volume of the stainless steel tubing con-nected from the desorption device to valve E, the moreconcentrated portion of the SFCO2 containing des-orbed pesticides could be adjusted to be in the 5-mlinjection loop. As the volume was changed from 42ml(B in Fig. 5) to 69ml (A in Fig. 5), larger peaks wereobtained. Therefore, the latter volume of the stainlesssteel tubing was used throughout the experiments.
As the upper limit of pressure of 350 atm advisedby the manufacturer could be used in the desorptionchamber, no attempt was made on the desorption ofthe pesticides above 306 atm for safety purposes. Con-sequently, only the desorption temperature was variedin this study. Rising the desorption temperature gener-ally decreases the density, hence the desorption abilityof SFCO2. In order to confirm this fact, the desorp-tion temperature was varied from 40 to 70◦C whilekeeping the desorption pressure constant at 306 atmand the results are shown in Fig. 6. As expected, thedesorption temperature of 70◦C produced lower peakarea for all pesticides while the desorption tempera-ture of 40 and 50◦C gave approximately equivalent
Fig. 6. Effect of the desorption temperature on the peak area ofpesticides desorbed from SPME fiber coating. Concentration ofpesticides: 1.0 mg l−1. Extraction conditions — time: 60 min andtemperature; 75◦C. Desorption conditions — pressure: 306 atm;temperature: 40, 50 and 70◦C; static desorption: 30 min and SFCO2
injection volume: 5ml.
S.H. Salleh et al. / Analytica Chimica Acta 433 (2001) 207–215 213
Fig. 7. Effect of static desorption time on the carryover of pesticideson the SPME fiber coating. Concentration of pesticides: 1.0 mg l−1.Extraction conditions — time: 60 min and temperature: 75◦C.Desorption conditions — pressure: 306 atm; temperature: 50◦Cand SFCO2 injection volume: 5ml.
peak area. Therefore, 50◦C was chosen as the opti-mum desorption temperature and will be used for allsubsequent experiments.
Fig. 7 shows the carryover versus static desorp-tion time profiles for the analytes. After the SFCO2containing desorbed pesticides was injected into theHPLC system, the fiber was desorbed again for 30 minwithout re-exposing to the spiked water solution. Thecarryover was calculated as the ratio of the peak areaof the pesticides remaining on the fiber after the firstdesorption to the total peak area of pesticides ex-tracted. The desorption time of 30 min was selected asthe optimum desorption time as most of the pesticidesreached 0% carryover although the carryover of EPNdid not decrease even after 60 min desorption. The op-timum desorption conditions obtained from this studyare summarized in Table 1.
Fig. 8 shows the chromatograms for the separationof 1 mg l−1 organophosphorus pesticides with the op-timized SPME/SFE/HPLC technique (A) and with di-rect injection (B). A shift in the retention time couldbe seen in the chromatogram, which may be attributedto the change of the viscosity of the mobile phase. Aportion of the mobile phase with lower viscosity willbe produced as the SFCO2 containing desorbed ana-lytes dissolved in the mobile phase. Since this portion
Fig. 8. Chromatograms of the pesticides extracted, desorbed, andinjected on-line into the SPME/SFE/HPLC system under the op-timized conditions (A) and the standard solution of pesticides in-jected directly into the HPLC system (B). The standard solutionis 1 mg l−1 each of pesticides. (1): Bensulide; (2): diazinon; (3):EPN; (4): chlorpyrifos.
of the mobile phase has a lower viscosity, it will elutethe analytes somewhat a little faster, giving the shiftin the retention time.
3.3. Limit of detection (LOD) and % R.S.D.s
For the determination of LOD, spiked water solu-tions were extracted and desorbed using the optimizedconditions listed in Table 1. In this case, LOD wasconsidered as the concentration that gave a signal tonoise ratio 3. The LOD for each compound used in thisstudy was found to be high (Table 2). The relatively
Table 2Method's limit of detection (LOD) and % R.S.D.s
Compound LOD (mg l−1) % R.S.D.s (n = 5)
Bensulide 600 9.6Diazinon 300 13.5EPN 40 11.3Chlorpyrifos 60 8.6
214 S.H. Salleh et al. / Analytica Chimica Acta 433 (2001) 207–215
Fig. 9. Chromatogram of a river water sample spiked with0.1 mg l−1 EPN and chlorpyrifos obtained with the optimizedSPME/SFE/HPLC technique. (1): EPN; (2): chlorpyrifos.
high detection limits are likely due to that only a por-tion of the SFCO2 containing desorbed pesticides, i.e.5ml was injected into the HPLC system. Injection ofhigher volume of SFCO2 could lower the LOD. In theother hand, in-tube SPME technique which has beenshown to give a very low detection limits for pheny-lurea pesticides [13,14,26], could also be adopted intothis system or more sensitive detection method suchas MS [11] could be used to achieve lower detectionlimits.
Table 2 also shows the accuracy of the developedmethod. Accuracy was determined by analyzing fivereplicate samples consecutively. The % R.S.D.s valuesobtained were a little poor compared to SPME/GC andSPME/HPLC with organic solvents as the desorptionsolvent. Since the density of SFCO2 change dramati-cally with temperature, keeping the 5-ml injection loopat a constant temperature might lower the % R.S.D.s.
3.4. Application to environmental water sample
The optimized SPME/SFE/HPLC was applied to awater sample taken from Umeda River, Toyohashi,Japan on 1 October 2000. The river water sample
Table 3Recoveries of pesticides spiked in river water samplea
Pesticides Recovery (% R.S.D.s,n = 3)
EPN 64 (10.2)Chlorpyrifos 62 (11.6)
a The values for suspended solid (SS), total organic carbon(TOC) and chemical oxygen demand (COD) of the river watersample are 9.6, 0.998 and 5.8 mg l−1, respectively.
was analyzed without any modification for the samplematrix. Since none of the pesticides was detected inthe water sample, 0.1 mg l−1 of EPN and chlorpyrifoswere spiked into the river water sample. Fig. 9 andTable 3 shows the chromatogram and the recovery ofthe analysis for the spiked river water sample, respec-tively. Recovery was calculated by dividing the peakareas of 0.1 mg l−1 pesticides spiked in the river wa-ter sample with that of a spiked standard water solu-tion. The values of suspended solid (SS), total organiccarbon (TOC) and chemical oxygen demand (COD)are also seen in Table 3. Taking into account the in-terference of the matrix, i.e. relatively high SS, morethan 50% of the pesticides could be analyzed with theSPME/SFE/HPLC technique showing good tendencytowards the application of this technique on environ-mental water samples.
4. Conclusion
In conclusion, the optimized desorption conditionsof pesticides using on-line injection of SFCO2 intothe HPLC system are as follows — pressure: 306 atm;temperature: 50◦C; static desorption time: 30 min andSFCO2 injection volume of 5ml. For the extractionconditions of pesticides from spiked water solution,temperature of 75◦C, and extraction time of 60 minwere found as the optimum conditions. The on-lineintroduction of SFCO2 containing desorbed pesti-cides into the HPLC system was also made success-fully eliminating the use of organic solvents and animpactor interface from the sample preparation pro-cedure although a relatively high LOD was obtainedfor each pesticide. This could be solved by employ-ing more sensitive detection method such as MS orinjecting larger volume of SFCO2 into the HPLCsystem. The SPME/SFE/HPLC technique was also
S.H. Salleh et al. / Analytica Chimica Acta 433 (2001) 207–215 215
successfully applied to the analysis of pesticidesspiked in a river water sample with the recoveries ofEPN and chlorpyrifos as 64 and 62%, respectively.
Acknowledgements
S.H. Salleh would like to thank Rabi'Atul'AdawiahOthman, School of Ecological Engineering, Toy-ohashi University of Technology, for providing valu-able help. The financial support by Japan Society forthe Promotion of Science Grant-in-Aid for ScientificResearch (C, No. 12640586) is acknowledged.
References
[1] Z. Zhang, J. Pawliszyn, Anal. Chem. 64 (1992) 1960.[2] K.D. Butchholtz, J. Pawliszyn, Environ. Sci. Technol. 27
(1993) 2844.[3] K.D. Butchholtz, J. Pawliszyn, Anal. Chem. 66 (1994) 160.[4] C.L. Arthur, D.W. Potter, K.D. Butchholtz, S. Motlagh, J.
Pawliszyn, LC–GC 10 (1992) 656.[5] C.L. Arthur, K. Pratt, S. Motlagh, J. Pawliszyn, J. High
Resolut. Chromatogr. 15 (1992) 741.[6] X.P. Lee, T. Kumazawa, K. Sato, O. Suzuki, Chromatographia
42 (1996) 135.[7] I.J. Barnabas, J.R. Dean, I.A. Fowlis, S.P. Owen, J.
Chromatogr. A 705 (1995) 305.
[8] K. Jinno, T. Muramatsu, Y. Saito, Y. Kiso, S. Magdic, J.Pawlizyn, J. Chromatogr. A 754 (1996) 137.
[9] R. Eisert, J. Pawliszyn, Anal. Chem. 69 (1997) 3140.[10] R. Eisert, J. Pawliszyn, Crit. Rev. Anal. Chem. 27 (1997) 103.[11] M. Möder, P. Popp, R. Eisert, J. Pawliszyn, Fresenius J. Anal.
Chem. 363 (1999) 680.[12] H. Hartmann, J. Burhenne, M. Spiteller, Fresenius Environ.
Bull. 7 (1998) 96.[13] Y. Saito, M. Kawazoe, M. Hayashida, K. Jinno, The Analyst
125 (2000) 807.[14] Y. Saito, Y. Nakao, M. Imaizumi, T. Takeichi, Y. Kiso, K.
Jinno, Fresenius J. Anal. Chem. 368 (2000) 641.[15] V. Camel, A. Tambute, M. Caude, J. Chromatogr. 642 (1993)
263.[16] S.B. Hawthorne, D.J. Miller, M.D. Buford, J.J. Langenfeld,
S. Eckert-Tilotta, P.K. Louie, J. Chromatogr. 642 (1993) 301.[17] T.L. Chester, J.D. Pinkston, D.E. Raynie, Anal. Chem. 66
(1994) 160R.[18] F.M. Lanças, M.A. Ruggiero, J. Microcol. Sep. 12 (2000) 61.[19] Y. Hirata, Y. Okamoto, J. Microcol. Sep. 1 (1989) 46.[20] M.H. Liu, S. Kapila, S. Nam, A.A. Elseewi, J. Chromatogr.
639 (1993) 151.[21] C. Mougin, J. Dubroca, E. Barriuso, J. High Resolut.
Chromatogr. 19 (1996) 700.[22] M. Asharaf-Khorassani, M. Barzegar, Y. Yamini, J. High
Resolut. Chromatogr. 18 (1995) 472.[23] S.H. Salleh, Y. Saito, K. Jinno, Chromatography 20 (1999)
126.[24] S.H. Salleh, Y. Saito, K. Jinno, Anal. Chim. Acta 418 (2000)
69.[25] M.W. Francis, J. Phys. Chem. 58 (1954) 1099.[26] R. Eisert, J. Pawliszyn, Anal. Chem. 69 (1997) 3140.
