957 (2002) 227–234Journal of Chromatography A,www.elsevier.com/ locate /chroma

D ynamic sonication-assisted solvent extraction of organophosphateesters in air samples

a a a,b a˚ ¨Cristina Sanchez , Magnus Ericsson , Hakan Carlsson , Anders Colmsjo ,a ,*Eva Dyremark

aDepartment of Analytical Chemistry, Stockholm University, SE-106 91 Stockholm, Swedenb ¨Swedish Defence Research Agency, FOI, Weapons and Protection Division Grindsjon, SE-147 25 Tumba, Sweden

Received 15 January 2002; received in revised form 11 March 2002; accepted 13 March 2002

Abstract

A new system for extracting solid samples based on dynamic sonication-assisted solvent extraction (DSASE) is described.The technique is highly efficient with respect to both time and solvent consumption. In tests reported here, organophosphateesters were extracted from air sampling filters in 3 min with an extraction volume of 600 mL of solvent. Furthermore, it waspossible to replace a previously used chlorinated solvent with a halogen-free solvent mixture. The sample was placed in acartridge through which fresh solvent was pumped continuously. A restrictor connected to the outlet of the cartridge allowedthe system to be used at a temperature of 70 8C without reaching the boiling point of the solvent. Both spiked and non-spikednative samples were used for the evaluation, which clearly revealed a stronger analyte–matrix interaction in native samples.The DSASE technique was shown to recover larger amounts of organophosphate esters from native samples, compared to astatic method. DSASE was applied to air samples collected in a lecture hall and from above a computer monitor. 2002Elsevier Science B.V. All rights reserved.

Keywords: Dynamic sonication-assisted solvent extraction; Sonication-assisted solvent extraction; Extraction methods; Airanalysis; Environmental analysis; Organophosphate esters; Organophosphorus compounds; Phosphate

1 . Introduction [6,7] and supercritical fluid extraction (SFE) [8].However, other techniques such as Soxhlet extrac-

The analysis of solid samples, including air sam- tion [9] and sonication [10–14] are still widely used.pling adsorbents and filters, in most cases involves In sonication, acoustic vibrations with frequenciesan extraction with a solvent [1,2]. There has been above 20 kHz are applied to the sample. When theseincreasing interest recently in the development of vibrations are transmitted through the liquid, cavita-various techniques for the extraction of solid sam- tion occurs, i.e. bubbles with a negative pressure areples, such as pressurized liquid extraction (PLE) formed. Chemical compounds and particles are re-[3–5], microwave-assisted solvent extraction (MAE) moved mechanically from the matrix surface and

adsorption sites by the shock waves generated whenthe cavitation bubbles collapse [15]. Further, implo-*Corresponding author. Tel.: 146-8-162439; fax: 146-8-sion of cavities creates microenvironments with high156391.

E-mail address: eva.dyremark@anchem.su.se (E. Dyremark). temperatures and pressures. Because of these pro-

0021-9673/02/$ – see front matter 2002 Elsevier Science B.V. All rights reserved.PI I : S0021-9673( 02 )00318-7

957 (2002) 227–234228 C. Sanchez et al. / J. Chromatogr. A

cesses, sonication can be used to decompose or 2 . Experimentaloxidize organic compounds. Hence, when developingan extraction method, care must be taken to avoid 2 .1. Components and use of the dynamicdegradation of the analytes. Dynamic extraction can sonication-assisted liquid extractorbe advantageous in this respect, since the analytesare removed as soon as they are transferred from the Samples to be extracted were inserted into thesolid matrix to the solvent. Further, in a dynamic extraction cell, an in-line preparative refillable guardsystem the sample is continuously exposed to fresh column (Alltech, Deerfield, IL, USA) with an inter-solvent, which promotes the transfer of analytes from nal volume of 0.25 mL. Solvent was pumped throughthe sample matrix to the solvent. this cartridge by means of a Varian 9012 (Varian,

Sonication-assisted liquid extraction has previous- Walnut Creek, CA, USA) HPLC pump, and ex-ly been used successfully in static mode for many traction was performed inside a Bransonic 52 ul-compounds. For instance, pesticides have been re- trasonic bath, with an output power of 120 W and acovered from soil by ultrasonic extraction [16] and frequency of 35 kHz (Branson, Soest, Netherlands).trace elements in lyophilized seafood samples have A Tempunit TU-16A heater (Techne, Cambridge,been determined after acid leaching induced by UK) was inserted into the ultrasonic bath to regulateultrasonic bath treatment [17]. Experimental design the temperature. An outlet restrictor made of fusedexperiments showed that, in order to extract silica (15 mm350 mm I.D., from Micro-Tech Sci-selenium from mussel tissue, high sonication tem- entific, Sunnyvale, CA, USA) was coupled to theperatures and long sonication exposure times are outlet tubing, in order to keep the solvent in therequired [17], while a factorial design experiment liquid state. The extracts were collected in glasshas shown that sample and solvent volumes are vials. A schematic diagram of the apparatus is showncritical parameters in the static ultrasonic extraction in Fig. 1.of volatiles in wine [18]. Carlsson et al. have A factorial design was used in experiments toreported recoveries higher than 95% in static investigate the effects of temperature, flow-rate andsonication-assisted extraction with dichloromethane duration of extraction. For this study, the data werefor determination of organophosphate esters spiked processed using Modde 4.0 software (Umetrics,

˚on an air sampling filter [19]. Umea, Sweden), a Windows-based program forOrganophosphate esters are used as flame retar- designing experiments. The order of the experiments

dants and plasticisers in a wide range of products. was not randomized; instead they were performed inThey are used as additives that are not covalently the order of increasing temperature. No effect on thebonded to the main material, e.g. plastics, and hence recovery or systematic error due to the non-random-may migrate out of the product [20–22]. Several ized order was detected.biological effects of exposure to organophosphateesters have been reported, including contact allergy 2 .2. Chemicals and standardsto triphenyl phosphate [23], which has also beenshown to be a powerful inhibitor of the monocyte Triethyl, tri(n-propyl), tri(n-butyl), tri(2-chloro-carboxylase in human blood [24]. In studies of rats ethyl), triphenyl, tri(2-butoxyethyl), triethylhexyl,and mice, tri(2-chloroethyl) phosphate has shown and tritolyl phosphates were obtained from Aldrichneurotoxic and carcinogenic properties [25]. Further- (Germany). Tri(chloropropyl) phosphate was pro-more, tri(2-chloroethyl) phosphate and tri(chloro- vided by Akzo Nobel (Sweden). All reference sub-propyl) phosphate have exhibited gonadotoxic ef- stances were of analytical grade (.98%) except thefects when tested on rats [26,27]. methyldiphenyl phosphate, which was of technical

This study reports further developments in grade (80%). Hence, this substance was purified bysonication-assisted liquid extraction, namely the semipreparative high-performance liquid chromatog-construction of a dynamic system and the evaluation raphy (HPLC) with an octadecylsilica column

¨of its suitability for extracting organophosphate (Macherey–Nagel, Duren, Germany). For develop-esters from air sampling filters. ment of the method, filters were spiked with 10 mL

957 (2002) 227–234 229C. Sanchez et al. / J. Chromatogr. A

Fig. 1. Schematic diagram of the dynamic extraction system: 15solvent, 25pump, 35extraction cell, 45ultrasonic bath, 55heater, 65zerodead-volume connector, 75restrictor, 85vial, 95GC–TID.

of a standard solution containing nine organophos- solvent used for the extraction in an ultrasonic bathphate esters with a concentration of approximately to eliminate particulate material from the filters0.4 ng/mL of each compound. A solution of 0.5 retained in the frits. Prior to the next extraction, freshng/mL of methyldiphenyl phosphate (Aldrich) was solvent was pumped through the system for 5 min atused as volumetric standard. Prior to GC analysis, 10 a flow-rate of 0.5 mL/min.mL of this solution was added to the sample extracts. Due to the ubiquity of organophosphates as indoorThe external standard was prepared by adding 10 mL air pollutants, materials used during sampling andof the standard solution and 10 mL of the volumetric sample handling require extensive cleaning [19].standard solution to 100 mL of the solvents used for Therefore, before use all glassware was soaked in athe respective extractions. solution of 5% (w/w) sodium hydroxide in ethanol

All standard substances were dissolved in either for 12 h and afterwards rinsed carefully with water,hexane (Merck, Darmstadt, Germany) or dichloro- ethanol and acetone. The glass fiber filters weremethane (Riedel-de Haen, Seelze, Germany). The sonicated for 20 min in methanol, acetone andfollowing solvents were tested as extraction solvents: dichloromethane, respectively.methyl tert.-butyl ether (MTBE) (Rathburn, Walk-erburn, UK), dichloromethane (Riedel-de Haen), 2 .4. Static extractionhexane and toluene (Merck). All solvents were ofanalytical grade. In order to evaluate the dynamic extraction tech-

nique described in this work, it was compared to a2 .3. Sample handling and clean-up procedure static method, in which the samples were extracted

in an ultrasonic bath without temperature regulationThe spiked filters were kept inside a box that had using 235 ml hexane–MTBE (7:3) for 2320 min.

previously been rinsed with dichloromethane, cov- The combined extract was then filtered throughered with aluminum foil for 30 min prior to ex- densely packed glass wool. All other operations weretraction in order to allow the solvent carrying the performed in the same way for both methods.spiked compounds to evaporate. The sample filterswere cut into small pieces and inserted into the 2 .5. Gas chromatographyextraction cell. After extraction, the collected solventwas reduced at ambient temperature to approximate- GC analysis was performed using an 8000 Top gasly 100 mL, using a gentle stream of nitrogen. chromatograph (Carlo Erba, Milan, Italy) equipped

After every extraction, the cartridge was removed with a DB-5MS column (30 m30.32 mm I.D., 0.10from the holder and cleaned for 5 min with the mm film thickness; J&W Scientific, Folsom, CA,

957 (2002) 227–234230 C. Sanchez et al. / J. Chromatogr. A

USA) and a TS-2 nitrogen–phosphorus detection filters [19]. For environmental and legislative(thermionic detection, TID) system. Nitrogen was reasons, in this study the possibility was tested ofused as both carrier and make-up gas. The column using a non-halogenated solvent to extract or-oven temperature was programmed as follows: 50 8C ganophosphate esters by dynamic extraction.held isothermally for 2 min during injection, fol- Organophosphate esters are not an homogeneouslowed by a temperature increase of 40 8C/min to 190 group with respect to chemical and physical prop-8C, 10 8C/min to 230 8C and 40 8C/min to 300 8C, erties, since there are wide-ranging structural varia-which was maintained for 10 min. Samples were tions in their substituents. Due to the huge amount ofinjected in splitless mode and the split was opened 2 experimental data obtained in this study, an isomermin after injection. Both detector and injector tem- of tri(2-chloropropyl) phosphate, triphenyl phosphateperatures were set to 300 8C. Chromatographic data and tri(2-ethylhexyl) phosphate were chosen towere registered and processed by ELDS Win Pro v represent the whole set of analytes. A property that is1.1 software (Chromatography Data System, commonly used to describe solubility properties is

¨Svartsjo, Sweden). the octanol–water partition coefficient (K ), ex-ow

pressed as the logarithm, log K The log Kow. ow

2 .6. Air sampling values of trichloropropyl phosphate, triphenyl phos-phate and tri(2-ethylhexyl) phosphate are 2.50, 4.70

Organophosphate esters were collected on 25 mm and 9.49, respectively [29]. Thus, the entire lipo-binder-free A/E borosilicate glass fiber filters (Gel- philicity range for these analytes was covered.man Science, Ann Arbor, MI, USA). These filters Initially, toluene, hexane and MTBE were testedretain 99.98% of particles with a diameter larger than as single solvents for extraction of spiked samples.0.3 mm from aerosols at a flow-rate of 3 L/min. The However, in order to obtain high recoveries of thesampler holder was made of anodized aluminum wide range of organophosphate esters tested a sol-[28]. If gaseous semi-volatile material is also to be vent mixture was needed, so the extraction efficiencycollected, the sampler can be used with two 15315 of hexane–MTBE in different proportions was in-mm cylindrical polyurethane foam (PUF) plugs in vestigated (Fig. 2). The best recoveries, more thanseries behind the filter. However, no PUF adsorbents 95%, for all nine of the tested compounds werewere analyzed in this study since organophosphate obtained when a mixture of hexane–MTBE (7:3)esters have previously been shown to be retained was used.only in the filters [19]. Air was pumped through the Regarding the size of the extraction cell, it wassampler using a battery-operated personal sampler observed that the smallest internal volume, 0.25 mL,pump. The flow-rate was set to 1.0 L/min and gave the highest extraction efficiency, regardless ofsamples were collected for 480 min, giving a total the type of solvent. Furthermore, a small internal

3sampled air volume of 0.48 m . The samples were volume was crucial in order to ensure high repro-covered with aluminum foil and stored in a freezer ducibility. Irreproducible results were obtained whenuntil analysis. Air samples were collected in a new an excessively large extraction cell was used, in thislecture hall and from above a new computer monitor. case 3.0 mL, most likely due to non-homogenousAnalysis of blank samples showed no traces of flow distribution within the cell. In the large cellorganophosphate esters. filter pieces could presumably be dispersed to sectors

that were poorly swept by the solvent stream, whichmainly passes through areas with low flow resist-

3 . Results and discussion ance. Thus, the smallest cartridge that could hold thesamples was selected for further tests.

3 .1. Choice of solvent3 .2. Investigation of extraction parameters using

A previously used, classical ultrasonic extraction factorial designtechnique gave recoveries higher than 95% usingdichloromethane as solvent, when applied to spiked There are several variables that can affect the

957 (2002) 227–234 231C. Sanchez et al. / J. Chromatogr. A

Fig. 2. Recoveries of tri(2-chloropropyl), triphenyl and tri(2-ethylhexyl) phosphates using different extraction solvents. Temperature, 70 8C;flow-rate, 1 mL/min; duration of extraction, 10 min. The presented values are the mean values of three replicate extractions.

extraction process. They can be divided into continu- were: temperature, 708C; extraction duration, 3 min;ous and discrete variables. The extraction solvent is flow, 200 mL/min, yielding an extraction volume asthe most important discrete variable and a factorial low as 600 mL.design was carried out using the solvent mixture thathad given the highest recoveries in the first experi- 3 .3. Validation using non-spiked native samplesments, i.e. hexane–MTBE (7:3). Furthermore, theultrasonic frequency was fixed from the beginning as Five air samples were collected from an environ-all the experiments were performed with the same ment containing electronic equipment, in order toultrasonic bath, using an ultrasonic frequency of 35 compare DSASE and static extraction. These meth-kHz. The continuous variables tested were flow-rate, ods were also compared to DSASE performed asextraction duration and temperature. above but with the sonication switched off, i.e.

In order to investigate the effects of the continuous dynamic solvent extraction at increased pressurevariables and their interactions, a two-level factorial allowing heating of the solvent above its boilingdesign with center points was applied to dynamic point at 1 atm (1 atm5101 325 Pa). In order tosonication-assisted solvent extraction (DSASE) of compare the three techniques, each filter was cut intofilters spiked with a standard solution. Eight cube several pieces as small as possible, which wereand three center point experiments were performed. divided into three sets that were each extracted byAll the experiments were carried out twice, and theresults used for optimization were the average of the Table 1two replicates. Independent studies for all the com- Investigation of DSASE by means of a two-level factorial design

experiment with center points. The table shows the chosen valuespounds were carried out.for the continuous variables, and the best values found withThe low, central and high values chosen for therespect primarily to recovery, and secondarily to time and solventcontinuous variables, as well as the best conditionsconsumption

found, are shown in Table 1. Within the experimentalFactor Low Center High Selecteddomain, temperature was the only variable that hadFlow (mL/min) 200 600 1000 200clear effects on recovery. Hence, in order to keep theTime (min) 3 6.5 10 3time and solvent consumption low, the best combina-Temperature (8C) 40 55 70 70tion of conditions within the experimental domain

957 (2002) 227–234232 C. Sanchez et al. / J. Chromatogr. A

one of the methods. The proportion of the com- organophosphate esters were obtained in the secondpounds in each group of pieces was controlled by fraction.weight, assuming that the air sample was homoge-neously retained over the filter.

The results are shown in Fig. 3. A higher ef- 3 .4. Determination of organophosphate esters inficiency was obtained with DSASE compared to air samplesstatic sonication-assisted liquid extraction for all thetested organophosphate esters. The pressurized dy- Analyses of air samples collected from differentnamic heated solvent extraction was also shown to environments were performed using the DSASEbe an efficient technique for all analytes except method presented here. GC–TID chromatogramstriphenyl phosphate. These differences between the from analyses of samples collected in a new lecturethree extraction techniques were not observed for the hall and from above a new computer monitor areextraction of spiked filters, where the analytes are shown in Fig. 4. Organophosphate esters weremuch easier to desorb from the matrix. With spiked identified in all samples (Table 2). However, therefilters, the recoveries were close to 100% for all the were major differences in the pattern of organophos-analytes with all the investigated techniques. This phate esters found in the two locations. The samplesemphasizes the difficulties of validating extraction collected from above a computer monitor charac-techniques using spiked samples, and the advantage teristically had a high concentration of triphenylwhen standard reference materials are available. phosphate. Carlsson [21] has found that a source of

The relative standard deviation (RSD) of the triphenyl phosphate in computerized indoor environ-amounts extracted from the native non-spiked sam- ments is the plastic material in the outer cover ofples was 5–8% for DSASE, 4–12% when using computer monitors. They also reported that thepressurized hot dynamic solvent extraction, and 5– concentration of triphenyl phosphate decreases7% for static extraction. gradually with time. Our samples were taken from

When the DSASE extractions were prolonged, and just above a running computer monitor that had beenmore fractions were collected, less than 4% of the used for only a couple of days. The samples col-

Fig. 3. Comparison of extraction of non-spiked native air samples using different techniques. Hexane–MTBE (7:3) was used as solvent inall experiments. The results for DSASE were set to 1.0. Repeated DSASE extraction yielded additional organophosphate esterscorresponding to 4% of the total amount extracted. n 5 5.

957 (2002) 227–234 233C. Sanchez et al. / J. Chromatogr. A

lected in the lecture hall had high concentrations oftrichloropropyl phosphate. The organophosphate es-ters found in this study have previously been iden-tified in air samples collected in various types ofcommon indoor work environments [30].

4 . Conclusion

Dynamic sonication-assisted solvent extractionwas shown to recover organophosphate esters effi-ciently from glass fiber filters used in air sampling.DSASE is a fast and cost effective technique thatfacilitates the use of small volumes of non-halo-genated solvents. The small solvent volumes re-quired give DSASE the potential for online couplingto GC or HPLC, or for utilization in direct combina-tion with large volume GC injections. Pressurizeddynamic heated solvent extraction was also shown tobe a powerful technique for many compounds. Theneed for additional sonication depends on the ana-

Fig. 4. Chromatograms from analysis of air samples extracted lyte–matrix interaction. Sonication had no negativewith DSASE. (A) Lecture hall, (B) sample collected above a

effects on the recovery of any of the analytesrunning computer monitor, (C) standard mixture. Organophos-subjected to DSASE. In this study, where bothphate esters: 15triethyl, 25tripropyl, 35tributyl, 45tri(2-chloro-spiked and non-spiked native samples were used forethyl) phosphates, 5–75isomers of trichloropropyl phosphates,

85methyldiphenyl phosphate (volumetric standard), 95triphenyl, the evaluation, the risk of pitfalls when relying solely105tri(2-butoxyethyl), 115tri(2-ethylhexyl) phosphates, 12– on spiked samples was clearly shown, indicating the155isomers of tritolyl phosphate ester.

necessity of using native samples. Repeated extrac-

Table 2Concentrations of organophosphate esters found in air samples

3 3Substituent Elution Lecture hall (ng/m ) Above computer monitor (ng/m )order

Mean value SD Mean value SD

Tripropyl 2 2.48 0.15 nqTributyl 3 1.93 0.14 54.9 2.7Tri(2-chloroethyl) 4 47.9 2.4 7.55 0.38First isomer of

atrichloropropyl 5 306 15 5.77 0.29Second isomer of

atrichloropropyl 6 128 6.4 2.58 0.15Third isomer of

atrichloropropyl 7 17.9 1.2 0.314 0.025Tri(2-butoxyethyl) 8 3.27 0.23 8.14 0.41Triphenyl 9 3.18 0.19 560 28Tri(2-ethylhexyl) 10 0.346 0.028 nq

3SD, standard deviation, n 5 3; nq, detected but not quantified, value below 0.1 ng/ma Isomers of tri(chloro)propyl phosphate with unknown chlorine positions.

957 (2002) 227–234234 C. Sanchez et al. / J. Chromatogr. A

[13] I. Moret, R. Piazza, M. Benedetti, A. Gambaro, C. Barbante,tion of native samples, which in this case wasP. Cescon, Chemosphere 43 (2001) 559.performed with DSASE by collecting several frac-

[14] J. Wu, L. Lin, F. Chau, Ultrason. Sonochem. 8 (2001) 347.tions, provides a way to reduce the risk of in- ¨[15] A. Colmsjo, in: A. Neilson (Ed.), PAHs and Related Com-sufficient extraction. However, for the most adequate pounds, Springer, Berlin, 1998.method evaluation a standard reference material is [16] S. Basic, M. Petrolia, M. Kastelan-Macan, J. Chromatogr. A

823 (1998) 3.needed.˜ ˜[17] P. Bermejo-Barrera, O. Muniz-Naveiro, A. Moreda-Pineiro,

A. Bermejo-Barrera, Anal. Chim. Acta 439 (2001) 211.[18] D.H. Vila, F.J. Mira, M.A. Lucena, F. Recamales, Talanta 50

A cknowledgements (1999) 413.¨[19] H. Carlsson, U. Nilsson, G. Becker, C. Ostman, Environ. Sci.

Technol. 31 (1997) 2931.Ove Jonsson and Thorvald Staaf are acknowl-[20] Brominated Flame Retardants, Substance Flow Analysis andedged for purification of the methyldiphenyl phos-

Assessment of Alternatives, Alternative Flame Retardants,phate standard. The project was financially supportedDanish Environmental Protection Agency, Copenhagen,

by The Swedish Research Council for Environment, 2000.Agricultural Sciences and Spatial Planning, and The [21] H. Carlsson, Ph.D Thesis, Department of Analytical Chemis-Swedish National Institute for Working Life. try, Stockholm University, Stockholm, 1999.

[22] International Program on Chemical Safety, EnvironmentalHealth Criteria 192, Flame Retardants: A General Intro-duction, World Health Organization, Geneva, 1997.R eferences

[23] Environmental Health Criteria 111, Triphenylphosphate,International Program on Chemical Safety, World Health

¨[1] E. Bjorklund, T. Nilsson, Trends Anal. Chem. 19 (2000) Organization, Geneva, 1991.434. [24] E.A. Emmet, G. Lewis, F. Tanaka, M. Bleecker, R. Fox,

[2] J.R. Dean, G. Xiong, Trends Anal. Chem. 19 (2000) 553. A.C. Darlington, D.R. Synkowski, A.M.J. Dannenberg, W.J.[3] O. Heemken, N. Theobald, B. Wenclawiak, Anal. Chem. 69 Taylor, M.S.J. Levine, Occup. Med. 27 (1985) 905.

(1997) 2171. [25] H.B. Matthews, S.L. Eustis, J. Haseman, Fundam. Appl.¨[4] A. Kreisselmeier, H.W. Durbeck, J. Chromatogr. A 775 Toxicol. 20 (1993) 477.

(1997) 187. [26] N.R. Shepel’skaya, N.E. Dyshinevich, Khim. Prom-st. Ser.:[5] Y. Abrha, D. Raghavan, J. Hazard. Mater. B 80 (2000) 147. Toksikol. Sanit. Khim. Plastmass 4 (1979) 39.

¨[6] M. Ericsson, A. Colmsjo, J. Chromatogr. A 877 (2000) 141. [27] N.R. Shepel’skava, N.G. Dyshinevich, Gig. Sanit. 2 (1980)´ ´[7] L.E. Garcıa-Ayuso, J.L. Luque-Garcıa, M.D. Luque de 85.

Castro, Anal. Chem. 72 (2000) 3627. ¨ ˚ ¨[28] C. Ostman, H. Carlsson, A. Bemgard, A. Colmsjo, in: P.[8] S.B. Hawthorne, in: R.M. Smith (Ed.), Supercritical Fluids in Garrigues, M. Lamotte (Eds.), Polycyclic Aromatic Com-

Chromatography and Extraction, Elsevier, Amsterdam, 1997; pounds: Synthesis, Properties, Analytical Measurements,S.B. Hawthorne, J. Chromatogr. A 785 (1997). Occurrence and Biological Effects, Gordon & Breach,

´[9] M.D. Luque de Castro, L.E. Garcıa-Ayuso, Anal. Chim. Acta Yverdon, Switzerland, 1993, p. 485.10 (1998) 369. [29] Physical Properties Database, Syracuse Research Corpora-

[10] F. Sun, D. Littlejohm, M.D. Gibson, Anal. Chim. Acta 364 tion, Syracuse, NY.(1998) 1. ¨[30] H. Carlsson, U. Nilsson, C. Ostman, Environ. Sci. Technol.

[11] J. Wang, K. Ashley, D. Marlow, Anal. Chem. 71 (1999) 34 (2000) 3885.1027.

[12] C.C. Nascentes, M.A.Z. Korn, M. Arruda, Microchem. J. 69(2001) 37.