Analytica Chimica Acta 520 (2004) 193–200

Pressurized liquid extraction in determination of polychlorinated

biphenyls and organochlorine pesticides in fish samples
Petr Suchan, Jana Pulkrabová, Jana Hajšlová, Vladimı́r Kocourek∗
Department of Food Chemistry and Analysis, Faculty of Food and Biochemical Technology, University of Chemical Technology,
Technická 5, 16628 Prague 6, Czech Republic

Received 16 December 2003; received in revised form 24 February 2004; accepted 24 February 2004

Available online 26 April 2004

Abstract

Pressurized liquid extraction (PLE) is a relatively new technique applicable for the extraction of persistent organic pollutants from various ma-
trices. The main advantages of this method are short time and low consumption of extraction solvent. The effects of various operational parame-
ters (i.e. temperature of extraction, number of static cycles and extraction solvent mixtures) on the PLE efficiency were investigated in this study.
Fish muscle tissue containing 3.2% (w/w) lipids and native polychlorinated biphenyls (PCBs), organochlorine pesticides (OCPs) and other
related compounds was used for testing. Purification of crude extracts was carried out by gel permeation chromatography employing Bio-Beads
S-X3. Identification and quantitation of target indicator PCBs and OCPs was performed by high-resolution gas chromatography (HRGC) with
two parallel electron-capture detectors (ECDs). Results obtained by the optimized PLE procedure were compared with conventional Soxhlet
extraction (the same extraction solvent mixtures hexane–dichloromethane (1:1 v/v) and hexane–acetone (4:1 v/v) were used). The recoveries
obtained by PLE operated at 90–120 ◦ C were either comparable to “classic” Soxhlet extraction (for higher-chlorinated PCB congeners and
DDT group) or even better (for lower chlorinated analytes). The highest recoveries were obtained for three static 5 min extraction cycles.
© 2004 Elsevier B.V. All rights reserved.

Keywords: Polychlorinated biphenyls; Organochlorine pesticides; Accelerated solvent extraction; Soxhlet extraction; Environmental samples

1. Introduction octanol–water partition coefficient Kow ) and hydrophobic.

Due to the persistence and lipophilic properties of the
Persistent organic pollutants (POPs) are chemicals that POPs, they are able to accumulate in the ecosystem. In
may persist for a long period of time in the environment. addition to carcinogenetic/mutagenic potential they may
POPs include polychlorinated biphenyls (PCBs), polychlori- cause toxic effects on animal reproduction, development,
nated dibenzo-p-dioxins (PCDDs) and dibenzofurans (PC- and immunological function.
DFs), polycyclic aromatic hydrocarbons (PAHs), organo- The analyses of PCBs and OCPs play an important role in
chlorine pesticides (OCPs) such as 1,2,3,4,5,6-hexachloro- the monitoring of environmental contamination. Information
cyclohexane (HCH) mixed isomers, 1,1,1-trichloro-2,2-bis needed for regulatory purpose as well as ecotoxicological
(4-chlorophenyl)ethane (DDT) and its degradation products risk assessments is obtained in this way. It should be noted
DDE and DDD, chlorobenzenes, and several other anthro- that a good accuracy of generated data is a critical aspect in
pogenic compounds. POPs are already strictly regulated and any decision-making process.
most of them are not currently in production. These chem- Determination of PCBs and OCPs in biotic samples com-
icals are prone to long-range transport through the upper monly consists of three steps: (i) extraction, (ii) purifica-
levels of atmosphere and can be deposited 1000 miles away tion/fractionation of crude extract and (iii) chromatographic
from the pollution source. Through atmospheric deposition separation, identification/quantitation. Although selective
[1] and river inputs [2] they have spread to all aquatic envi- extraction of organics appears to be an attractive option,
ronments. POPs are characterized by being lipophilic (high different types of binding of analytes on to adsorption sites
or their interactions with matrix require an exhaustive tech-
∗ Corresponding author. Tel.: +420-2-24353185;
nique to recover the maximum amount of target analytes
fax: +420-2-24353185.
from the substrate. Generally, low selectivity of extrac-
E-mail address: [email protected] (V. Kocourek). tion method yields considerable amounts of undesirable

0003-2670/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.aca.2004.02.061
194 P. Suchan et al. / Analytica Chimica Acta 520 (2004) 193–200

co-extractives (e.g. lipids, pigments) which must be re- erational PLE parameters. This material is commonly used
moved before the final GC determination. Lipid separation in our laboratory as internal reference material in trace-
is usually performed by gel permeation chromatography ability to CRM 350 and CRM 430 (BCR, Belgium). The
(e.g. Bio-Beads) or adsorption chromatography (Florisil, average lipid content of this material is 3.16%. The fish
silica gel, alumina, etc.) [3–5]. tissue was chosen because fish products are considered as
The most common procedure employed for the extrac- significant source of PCBs in the diet. Some fish species are
tion of non-polar and semi-polar trace organics compounds commonly used in the environmental monitoring of these
(e.g. PCBs and OCPs) from a wide variety of matrices compounds.
such as sediments, soils, animal and plant tissues is Soxh-
let extraction [6–10]. This procedure is carried out with
2.2. Chemicals
non-polar and/or semi-polar solvents (pentane, hexane,
dichloromethane, acetone, and diethyl ether) or their mix-
The mixture of indicator PCBs (IUPAC numbers 28, 52,
tures advantageously in azeotrophic ratio. It should be
101, 118, 138, 153 and 180) in isooctane (trimethylpen-
noted that the Soxhlet method requires large volumes of
tane) and standards (solids) of organochlorine pesticides
highly purified organic solvents and a relatively long time
(hexachlorobenzene (HCB), ␣-, ␤- and ␥-isomers of hex-
for completing total extraction of all analytes. In case of
achlorocyclohexane, octachlorostyrene (OCS), DDT and its
high-moisture samples their thorough desiccation is needed
degradation products DDE and DDD) used in this study
to enable good penetration of solvent into the sample ma-
were obtained from Dr. Ehrenstorfer (Germany). The pu-
trix. In addition, some volatile compounds may be lost
rity of individual standards was higher than 96%. Working
unless efficient condensers are used [11].
standard solutions were prepared in isooctane and stored in
The other very simple isolation technique is batch extrac-
the refrigerator (5 ◦ C).
tion enhanced by sonication. “Dirty” extracts (containing
The solvents used (n-hexane, dichloromethane, isooctane,
a lot of co-extracted matrix components) are typically ob-
cyclohexane) were all supplied by Merck (Germany). Ethyl
tained by both the above-mentioned techniques. In the last
acetate was supplied by Scharlau (Spain). All these sol-
decade, alternative extraction techniques enabling reduction
vents were “organic trace analysis” grade. Acetone was
of the volume of extraction solvents and saving the extrac-
obtained from Lachema (Brno, Czech Republic) and re-
tion time have been searched. Supercritical fluid extraction
distilled before use. Anhydrous sodium sulphate obtained
(SFE), microwave-assisted extraction (MAE), and/or pres-
from Penta Chrudim was heated at 500 ◦ C for 5 h and then
surized liquid extraction (PLE; Dionex trade name ASE for
stored in the desiccators before use. Styrene–divinylbenzene
accelerated solvent extraction) represent recent techniques
gel (Bio-Beads S-X3, 200–400 mesh) was purchased from
meeting at least some of these requirements [12–17].
Bio-Rad (USA). Sulphuric acid (98%) was obtained from
The latter one, PLE uses conventional liquid solvents
Merck.
at elevated pressures (10–15 MPa) and temperatures (50–
200 ◦ C) for short time periods (5–10 min) to extract solid
samples quickly and with much less solvent than con- 2.3. Equipment
ventional techniques [18–21]. Nevertheless, samples with
high-moisture content require desiccation before the ex- A Dionex ASE 100 system (Dionex, USA) with stainless
traction step. Hexane, hexane–dichloromethane (1:1 v/v), steel vessels (66 ml) and Soxhlet extractor (Gerhardt 173200
hexane–acetone (1:1 v/v) and toluene are extraction solvents EV, Germany) with cellulose extraction thimbles (Whatman,
often used for the isolation of PCBs and OCPs from abi- UK) were used for realization of the extraction step.
otic (sediment, soil, sludge, urban dust) or biotic samples An automated gel permeation chromatography (GPC) sys-
(oyster, mussel) [21–25]. tem consisting of 305 MASTER pump, fraction collector,
The aim of this study was: (i) to optimize extraction effi- automatic regulator of loop XLI, microcomputer (software
ciency of PLE employing different extraction solvent mix- 731 PC via RS232C), dilutor 401C (Gilson, France) and
tures under the different extraction conditions and (ii) to stainless steel column 500 mm × 8 mm I.D. packed with
compare performance characteristics of this novel technique Bio-Beads S-X3, 200–400 mesh (Bio-Rad) was used for the
with conventional Soxhlet extraction. clean-up of extracts.
Vacuum evaporator (Büchi Rotavapor R-114 a Waterbath
B-480, Switzerland) was used for the concentration of ex-
2. Experimental tracts.
A Hewlett-Packard 5890 Series II gas chromato-
2.1. Sample material graph equipped with electronic pressure control (EPC),
split/splitless injector, two parallel 63 Ni electron-capture de-
The fish filets homogenate (skin removed) prepared from tection (ECD) systems and two parallel columns possessing
chub (Leuciscus cephalus) containing “native” levels of different selectivity (DB-5 and DB-17, both J&W Scientific,
PCBs and OCPs was used for the testing of individual op- USA) were employed for the analyses of PCBs and OCPs.
 P. Suchan et al. / Analytica Chimica Acta 520 (2004) 193–200 195

2.4. Analytical method as an internal standard (to our experience, this congener
never occurs in fish samples above detection limits).
2.4.1. Isolation
Two alternative procedures described below were used for 2.4.1.2. PLE. The sample preparation procedure was
isolation of analytes. carried out in the same way as in the case of Soxhlet ex-
traction, including desiccation. The flowing powder was
2.4.1.1. Soxhlet extraction. Samples with high-moisture placed into the extraction cell. The remaining volume of the
content have to be desiccated before the extraction step to cell was completely filled with anhydrous sodium sulphate.
enable good penetration of solvent into the sample matrix. The sample cell was loaded into the ASE 100 system.
The 10 g fish homogenate were mixed with 70 g anhydrous Hexane–dichloromethane (1:1 v/v) and hexane–acetone
sodium sulphate to form a flowing powder. This sample was (4:1 v/v) were used for testing of extraction solvents. The
then transferred into a cellulose extraction thimble and stored experimental set-up of PLE testing is shown in Fig. 1. The
in desiccators for 12 h. After this time it was inserted into pressure 10 MPa, flush volume 60% of the extraction cell
a Soxhlet apparatus and extracted for 8 h (7 cycles/h). The volume (the volume of fresh extraction solvent mixture
170 ml of solvent mixture, either hexane–dichloromethane used for flushing of the extraction cell after static extrac-
(1:1 v/v) or hexane–acetone (4:1 v/v), were used as an extrac- tion, i.e. about 40 ml) and purge time (N2 ) 5 s were used for
tion solvent. The crude extracts were carefully evaporated all PLE experiments. Processing of obtained crude extracts
by rotary vacuum evaporator at 40 ◦ C and the rest of solvent was identical as for Soxhlet extraction.
was removed by a gentle nitrogen stream. The lipid content
was determined gravimetrically. Extracted lipids were dis- 2.4.2. Clean-up
solved in 4.5 ml of cyclohexane–ethyl acetate (1:1 v/v) in- A clean-up of crude extracts was carried out by GPC
cluding PCB congener No. 112 in a concentration of 5 ng/ml employing Bio-Beads S-X3. Cyclohexane–ethyl acetate

Solvent extraction mixture

hexane-dichloromethane (1:1, v/v)

Changing extraction temperature Changing number of static cycles

Static extraction time: 5 min Temperature: 90˚C

Number of static cycles: 2 Tested number of static cycles:
Tested temperatures: 1 x 5 min 2 x 5 min 3 x 5 min
60˚C 80˚C 90˚C 100˚C 120˚C
(A)

Solvent extraction mixture

hexane-acetone (4:1, v/v)

Changing extraction temperature Changing number of static cycles

Static extraction time: 5 min Temperature: 90˚C

Number of static cycles: 2 Tested number of static cycles:
Tested temperatures: 2 x 5 min 3 x 5 min
90˚C 100˚C
(B)

Fig. 1. Experimental set-up: optimization of operational parameters for both tested extraction solvent mixtures: (A) hexane–dichloromethane (1:1 v/v) and
(B) hexane–acetone (4:1 v/v).
196 P. Suchan et al. / Analytica Chimica Acta 520 (2004) 193–200

(1:1 v/v) was used as mobile phase; the flow rate was Table 1
0.6 ml/min; injection volume 2 ml. The fraction correspond- GC conditions used for indicator PCBs and OCPs analysis
ing to the elution volume of 14.5–28.0 ml was collected. Parameter Description
The eluate was evaporated by rotary vacuum evaporator Column types DB-5 (5% phenyl–methylpolysiloxane),
at 40 ◦ C and the rest of solvent was carefully removed by DB-17 (50%
a gentle stream of nitrogen. The residue after evaporation phenyl–methylpolysiloxane)
was dissolved in 1 ml of isooctane and treated with concen- Column size 60 m × 0.25 mm I.D., 0.25 ␮m (both)
trated sulphuric acid (1 ml) to remove potentially remaining Injector temperature 250 ◦ C
Detector temperatures 300 ◦ C
lipids. Aliquot of upper isooctane layer was transferred into Oven temperature program 60 ◦ C (2.5 min), 30 ◦ C/min to 220 ◦ C,
a glass vial for further GC analysis. 0.5 ◦ C/min to 240 ◦ C, 2.5 ◦ C/min to
280 ◦ C (10 min)
2.4.3. GC/ECD determination Splitless period 2.0 min
HRGC/2ECD was used for analysis of PCBs and OCPs. Carrier gas Helium
Inlet pressure program Constant flow 1.7 ml/min, i.e.
GC conditions are summarized in Table 1. Identification of 207 kPa (60 ◦ C)
analytes was carried out by comparison of retention times in Linear velocity of carrier gas 29.3 cm/s
chromatogram with that of PCBs and OCPs standards. Quan- Injected sample volume 1 ␮l
titation of target compounds was performed by multi-level Data processing software HP GC ChemStation Rev.A.06.03
calibration. (509)

samples have been published, none of them focused on a

3. Results and discussion comparison of this technique with conventional approaches.
In this first part of experiments the attention was paid to a
3.1. Optimization of PLE parameters critical assessment of the influence of various parameters on
PLE performance and selection of optimal ones. Various ex-
Although several studies concerned with application of traction temperatures together with different number of static
PLE for the isolation of PCBs and/or OCPs from biotic cycles were tested. The extraction efficiency of two solvent

Fig. 2. The effect of extraction temperature on the extraction efficiency of PLE for: (A) indicator PCBs, lipids and (B) OCPs; extraction solvent
hexane–dichloromethane (1:1 v/v) (Soxhlet extraction with hexane–dichloromethane was set at 100%).
 P. Suchan et al. / Analytica Chimica Acta 520 (2004) 193–200 197

mixtures (hexane–dichloromethane and hexane–acetone) into matrix particles and enhance extraction efficiency [21].
was compared as well. The data obtained under tested pa- Fig. 2 shows the effect of temperature setting on the extrac-
rameters were compared with those achieved by Soxhlet tion efficiency of PLE for individual analytes from fish ho-
extraction that is commonly used in one laboratory for mogenate. Experiments were carried out at constant pressure
isolation of halogenated POPs from both biotic and abiotic (10 MPa). The most pronounced increase of PLE efficiency
matrices. was observed in the temperature range 60–90 ◦ C, further
All experiments were designed to eliminate random errors increasing of this parameter did not result in significantly
by means of replicates. Each result shown in bar graphs elevated recoveries. The extraction efficiency of PLE for
(Figs. 2–4) represents the average value calculated from at higher-chlorinated PCB congeners (Nos. 101, 118, 138, 153,
least five individual experiments/replicates and subjected to 180) and p,p -DDE, p,p -DDD and p,p -DDT in the range
t-test to determine the statistical differences between mean 90–120 ◦ C was comparable to “classic” Soxhlet extraction;
values. Typical relative standard deviations (RSD, %) are for lower chlorinated PCBs and HCB it was even higher. On
stated in Table 2 and Fig. 4. These values are very similar in the other hand, for isomers of HCH and OCS slightly lower
all experiments and depend rather on the concentration level results were obtained with PLE as compared to Soxhlet ex-
of analyte than on the technique or solvent mixture used (in traction. It should be noted that only very low levels of these
fact, the F-test did not reveal any difference at all). analytes were contained in examined fish samples hence the
values of uncertainty of measurement were relatively high.
3.1.1. Influence of extraction temperature
The extraction temperature via changing diffusion co- 3.1.2. Influence of number of static cycles
efficient has a distinct influence on the extraction kinetics The number of extraction cycles is another important
and therefore also on overall recoveries [26]. In general, parameter for achieving quantitative extraction. Björklund
increased temperature can disrupt the strong solute–matrix et al. [27] showed the importance of this parameter in
interactions caused by Van der Waals forces, hydrogen study concerned with extraction of PCBs from sediments
bonding or dipole attractions of the solute molecules and with different particle sizes. While the target analytes were
active sites in the matrix. Higher temperatures also decrease quantitatively extracted with hexane–acetone mixture from
the viscosity of solvents, thus allow their better penetration materials with small differences in particle size using a

Fig. 3. Influence of number of static cycles on the extraction efficiency of PLE for: (A) indicator PCBs, lipids and (B) OCPs (number 1 × 5 min was
set-up as 100%); H–D: hexane–dichloromethane; H–A: hexane–acetone.
198 P. Suchan et al. / Analytica Chimica Acta 520 (2004) 193–200

Fig. 4. Comparison of relative efficiency of optimized ASE and Soxhlet extraction: two tested extraction solvent mixtures (values obtained by Soxhlet
extraction with hexane–dichloromethane were set-up as 100%): (A) PCBs and lipids and (B) OCPs; error bars represent standard deviation of results
(n = 5).

Table 2
Contents of PCBs and OCPs (␮g kg−1 fresh tissue) and repeatability of measurements (n = 5) obtained by procedures involving: (i) PLE and (ii) Soxhlet
extraction (H–D: hexane–dichloromethane; H–A: hexane–acetone); RSD: relative standard deviation (%)
Analyte Soxhlet, H–D (1:1) PLE, H–D (1:1) Soxhlet, H–A (4:1) PLE, H–A (4:1)

␮g kg−1 RSD (%) ␮g kg−1 RSD (%) ␮g kg−1 RSD (%) ␮g kg−1 RSD (%)

PCB 28 41 3 45 4 40 5 48 6
PCB 52 39 3 43 4 38 5 48 7
PCB 101 24 3 22 4 23 7 24 10
PCB 118 18 4 19 5 17 3 19 5
PCB 138 47 3 43 5 45 8 44 8
PCB 153 101 4 98 7 99 5 102 5
PCB 180 57 3 57 8 55 4 55 4
HCB 8 2 9 6 8 14 10 6
␣-HCH 0.8 5 0.8 14 0.8 14 0.8 9
␤-HCH 1.0 12 0.9 4 0.8 5. 0.9 5
␥-HCH 0.7 4 0.6 7 0.7 14 0.7 7
OCS 0.3 5 0.3 9 0.3 6 0.3 11
p,p -DDE 68 4 72 4 65 8 74 2
p,p -DDD 8 12 9 3 7 11 10 3
p,p -DDT 4 3 4 5 4 5 5 2

single 5 min extraction step (less than 1% PCBs of total the importance of application of repeated extraction cy-
content were found in the second extract), other very in- cles for efficient isolation of OCPs from soil. In general,
homogeneous sediments retained as much as 14% of some many “optimal” extraction set-ups/solvent mixtures can be
congeners, which were then completely released during found in literature [21,24,27–30]. In our study five experi-
a second 5 min step. Poop et al. [28] also emphasized ments differing in the number of extraction cycles and used
 P. Suchan et al. / Analytica Chimica Acta 520 (2004) 193–200 199

solvents were employed. The results are summarized in ples at very low concentration levels such as HCH iso-
Fig. 3. Negligible differences were found between one and mers and OCS, without regard to the extraction technique
two static cycles. On average, three static cycles provided used.
maximum recoveries for all target analytes and for both Laboratory throughput is an important issue in any routine
tested solvent mixtures but the differences between two laboratory. The total extraction time needed in our experi-
and three static cycles were not statistical significant. It ments for one, two and three static cycles was 11, 17, and
should be noted that three static cycles prolong the total 24 min, respectively. Using the ASE 300 device (designed
extraction time (as well as the solvent consumption) with- for high sample capacity), automated extraction of batch of
out any significant effect. Thus, two cycles offer a practical 12 samples last approximately 3.5 h using two extraction
solution. static cycles. It is a relatively short time in comparison with
The consumption of extraction solvent depends on both the Soxhlet technique, which requires approximately 8 h for
the number of extraction cycles and the portion of fresh sol- the efficient extraction of organochlorine pollutants from bi-
vent used to rinse the cell after the static extraction step. The otic samples.
total solvent volume for one, two and three static cycles was In addition, the estimated cost per sample for “classic”
70, 85, and 105 ml, respectively, which is significantly less Soxhlet extraction is about US$ 27 comparing to US$ 14
as compared with common Soxhlet technique. for PLE. This cost is based on the assumption of processing
of 2000 samples per year; appropriate laboratory equipment
3.1.3. Influence of extraction solvent mixtures (including its depreciation) personnel costs and consumables
In general, physico-chemical properties such as boiling were considered.
point, polarity, specific density (influences a penetration
into the sample matrix) as well toxicity (makes a workplace
hazard) are facts considered when choosing extraction sol- 4. Conclusions
vent [25]. For evaluation of the extraction solvents on PLE
efficiency, fish samples were extracted by solvent mixtures The results generated in this study document that for ex-
hexane–dichloromethane (1:1 v/v) and hexane–acetone traction of indicator PCBs and some organochlorine pes-
(4:1 v/v). Considering the recoveries of target analytes, ticides from fish matrix the performance characteristics of
regardless the solvent mixture used for extraction, no sta- PLE are essentially equivalent or even better than those of
tistically significant differences were found between the classic Soxhlet procedure. The efficiency of extraction for
results obtained by optimized PLE (extraction temperature some more volatile (HCB) as well as semi-volatile com-
100 ◦ C, number of static cycles 2 × 5 min) and those ob- pounds (PCB 28, 52) obtained by PLE was slightly higher
tained by Soxhlet extraction, see Fig. 4. The same results as compared to the Soxhlet extraction. The main advantages
were obtained for extracted lipids. This fact is very impor- of PLE were low consumption of extraction solvents and
tant since the concentrations of lipophilic contaminants are short time needed for the realization of extraction step (av-
often standardized to the lipid content. erage solvent consumption was 85 ml and extraction time
Separation of PCBs from co-extracted lipids is necessary was 17 min per sample while to 170 ml and 8 h in case of
to obtain extracts that can be analysed by GC. Using PLE, Soxhlet). However, the purchase cost of this equipment is
selective extraction of PCBs can be achieved by mixing much higher compared to common Soxhlet or batch ex-
the sample with suitable sorbent retaining lipids. Alumina traction enhanced by sonication. Another limitation of PLE
and/or silica gel impregnated by sulphuric acid can be used application is the maximum amount of sample that can be
for this purpose. Decrease of turbidity as well as the reduced placed into the extraction thimble under experimental con-
content of matrix pigments in crude extracts obtained by dition we used, only 10 g sample (flowing powder obtained
PLE could be affected by optimization of polarity of used by other addition of sodium sulphate). This might be draw-
extraction solvent mixture. In this particular case, the polar- back namely in case of analysis of sample with very low
ity parameter of hexane–dichloromethane (P  = 1.6) was levels of target analytes.
slightly higher in comparison with hexane–acetone (P  =
1.1) [31].
Acknowledgements
3.1.4. Repeatability of PLE and Soxhlet extraction
Five replicate extractions of fish samples were performed This study was realized as a part of the project “The
employing both tested extraction solvent mixtures to calcu- Influence of Environmental Pollution on Contamination
late repeatability of PLE and Soxhlet extraction (expressed and Quality of Biotic Components of Ecosystem” financed
as RSD, %). The results are shown in Table 2. The repeata- by Ministry of the Environment of the Czech Republic
bility of PLE ranged from 3 to 14% for all the target an- (VaV 340/1/01); part of the funding was also obtained
alytes, which was comparable to performance of conven- from research project No. MSM 223300004 granted by
tional Soxhlet extraction procedure. Generally, the worst re- the Ministry of Education, Youth and Sports of the Czech
peatability was obtained for analytes occurring in real sam- Republic.
200 P. Suchan et al. / Analytica Chimica Acta 520 (2004) 193–200

References [16] S. Bøwadt, B. Johansson, S. Wunderli, M. Zennegg, L.F. Alencastro,

D. Grandjean, Anal. Chem. 67 (1995) 2424.
[17] B.E. Richter, J.L. Ezzell, W.D. Felix, K.A. Roberts, D.W. Later, Am.
[1] S.J. Eisenreich, B. Loonez, J.D. Thorton, Environ. Sci. Technol. 15
Lab. (February 1995) 24.
(1981) 30.
[18] J.R. Dean, Anal. Commun. 33 (1996) 191.
[2] P. Larson, L. Olka, S. Rydlink, B. Westoo, Can. J. Fish. Aquat. Sci.
[19] N. Saim, J.R. Dean, M.P. Abdullah, Z. Zakaria, Anal. Chem. 70
47 (1990) 746.
(1998) 420.
[3] J. de Boer, Chemosphere 17 (1988) 1803.
[20] J.A. Fisher, M.J. Scarlett, A.D. Stott, Environ. Sci. Technol. 31
[4] D.L. Stalling, R.C. Tindle, J.L. Johnson, J. Assoc. Off. Anal. Chem.
(1997) 1120.
55 (1972) 32.
[21] B.E. Richter, B.A. Jones, J.L. Ezzell, N.L. Porter, Anal. Chem. 68
[5] J.D. McKinney, L. Moore, A. Prokopet, D.B. Walters, J. Assoc. Off.
(1996) 1033.
Anal. Chem. 67 (1984) 122.
[22] Z. Abrha, D. Raghavan, J. Hazard. Mater. B 80 (2000) 147.
[6] P.G. Wester, F. van der Valk, Environ. Contam. Toxicol. 45 (1990) 6.
[23] B.E. Richter, J.L. Ezzell, D.E. Knowles, F. Hoefer, Chemosphere 34
[7] R.C. Hale, E. Bush, K. Gallagher, J.L. Gundersen, R.F. Mothershead,
(1997) 975.
J. Chromatogr. 539 (1991) 149.
[24] M.M. Schantz, J.J. Nichols, S.A. Wise, Anal. Chem. 69 (1997) 4210.
[8] F. van der Valk, P.G. Wester, Chemosphere 22 (1991) 57.
[25] A. Hubert, K.D. Wenzel, W. Engelwald, G. Shüürmann, Rev. Anal.
[9] C. Porte, D. Barceló, J. Albaigés, Chemosphere 24 (1992) 735.
Chem. 20 (2001) 101.
[10] D.E. Wells, I. Echarri, Int. J. Environ. Anal. Chem. 47 (1992) 75.
[26] F. Höfler, J. Ezzel, B. Richter, Labor Praxis 3 (1995) 62.
[11] P. Hess, J. de Boer, W.P. Coffino, P.E.G. Leonards, D.E. Wells, J.
[27] J. Björklund, S. Bøwadt, T. Nilsson, L. Mathiasson, J. Chromatogr.
Chromatogr. A 703 (1995) 417.
A 836 (1999) 285.
[12] E. Björklund, Ch. Von Holst, E. Anklam, Trends Anal. Chem. 21
[28] J. Poop, P. Keil, M. Möder, A. Paschke, U. Thuss, J. Chromatogr.
(2002) 39.
A 774 (1997) 203.
[13] E. Conte, R. Milani, G. Morali, F. Abbable, J. Chromatogr. A 765
[29] Application Note 322, Dionex, Sunnyvale, CA, 1996.
(1997) 121.
[30] Application Note 337, Dionex, Sunnyvale, CA, 1999.
[14] O.P. Heemken, N. Theobald, B.W. Wenclawiak, Anal. Chem. 69
[31] V.R. Meyer, Practical High-performance Liquid Chromatography,
(1997) 2171.
2nd ed., Frankfurt/Main, 1996.
[15] S.P. Frost, J.R. Dean, K.P. Evans, K. Harradine, C. Cary, M.H.I.
Comber, Analyst 122 (1997) 895.