Talanta 78 (2009) 936–941

Contents lists available at ScienceDirect

Talanta
journal homepage: www.elsevier.com/locate/talanta

Separation and determination of benzene, toluene, ethylbenzene and o-xylene

compounds in water using directly suspended droplet microextraction coupled
with gas chromatography-flame ionization detector
A. Sarafraz-Yazdi a,c,∗ , A.H. Amiri a , Z. Es’haghi b,c
a
Department of Chemistry, Faculty of Sciences, Ferdowsi University of Mashhad, Iran
b
Department of Chemistry, Faculty of Sciences, Payame Noor University, Iran
c
Biotechnology Research Center of Ferdowsi University, Mashhad, Iran

a r t i c l e i n f o a b s t r a c t

Article history: The directly suspended droplet microextraction (DSDME) technique coupled with the capillary gas
Received 27 November 2008 chromatography-flame ionization detector (GC-FID) was used to determine BTEX compounds in aque-
Received in revised form ous samples. The effective parameters such as organic solvent, extraction time, microdroplet volume,
28 December 2008
salt effect and stirring speed were optimized. The performance of the proposed technique was evalu-
Accepted 30 December 2008
ated for the determination of BTEX compounds in natural water samples. Under the optimal conditions
Available online 20 January 2009
the enrichment factors ranged from 142.68 to 312.13, linear range; 0.01–20 ␮g mL−1 , limits of detection;
0.8–7 ng mL−1 for most analytes. Relative standard deviations for 0.2 ␮g mL−1 of BTEX in water were in the
Keywords:
Directly suspended droplet microextraction
range 1.81–2.47% (n = 5). The relative recoveries of BTEX from surface water at spiking level of 0.2 ␮g mL−1
(DSDME) technique were in the range of 89.87–98.62%.
BTEX © 2009 Published by Elsevier B.V.
Water sample
GC-FID

1. Introduction Historically, liquid–liquid extraction (LLE) and solid-phase

extraction (SPE) were often used for the extraction of haz-
Volatile organic compounds (VOCs) are organic chemical com- ardous compounds from aqueous matrices. However, LLE is
pounds that have high enough vapor pressures under normal time-consuming, generally labor-intensive, and requires large
conditions to significantly vaporize and enter the atmosphere. The quantities of expensive, toxic, and environmentally unfriendly
acronym BTEX defines the mixture of benzene, toluene, ethylben- organic solvents.
zene and the three xylenes isomers (ortho, meta and para), all being In case of SPE, in comparison with LLE, although it requires a
harmful VOCs. The effects of exposure to these substances comprise smaller volume of toxic organic solvents for the analyte desorption,
changes in the liver and damaging effects on the kidneys, heart, the small columns or disks used cause “plugging” if the aqueous
lungs, and the nervous system [1]. BTEX are emitted to the envi- sample contain fine solid particles. Therefore, nowadays the sim-
ronment from an extensive variety of sources including combustion plification and miniaturization of the sample preparation methods
products of wood and fuels, industrial paints, adhesives, degreas- are recommended, which have the advantages of either none or
ing agents and aerosols [2]. Therefore, they are ubiquitous among very little amount (␮L) of toxic organic solvents used.
samples of environmental concern (air, water and soil), the human Currently, the technique of Solid Phase Microextraction (SPME)
exposure to these aromatic hydrocarbons occurring by ingestion is used by some researchers. This technique allows a rapid and
(consuming contaminated water or food), inhalation or absorption solvent-free extraction of organic compounds from aqueous sam-
through the skin. In order to reduce the human intake of these haz- ples by partitioning between the stationary phase and the aqueous
ardous substances a chemical control (and consequently methods medium [3,4]. The technique is commercially available, and is capa-
of analysis) is desirable. ble of extracting micropollutants in aqueous samples prior to GC
analysis. The main drawbacks of SPME are (i) increase in the cost
of analysis per sample due to the use of certain special expen-
∗ Corresponding author at: Department of Chemistry, Faculty of Sciences, Fer-
sive apparatus, (ii) degradation of fibers with increased usage, and
dowsi University of Mashhad, Mashhad, Khorasan 91775, Iran. Tel.: +98 511 8432023;
(iii) carry-over between extractions [5]. In order to overcome these
fax: +98 511 8438032. problems, a simple and inexpensive Liquid-Phase Microextraction
E-mail address: asyazdi@ferdowsi.um.ac.ir (A. Sarafraz-Yazdi). (LPME) has been recently introduced.

0039-9140/$ – see front matter © 2009 Published by Elsevier B.V.

doi:10.1016/j.talanta.2008.12.069
 A. Sarafraz-Yazdi et al. / Talanta 78 (2009) 936–941 937

In this technique, only several microliters of solvents are easily synthesized or commercially available. These characteristics
required to concentrate analytes from aqueous samples rather than have led to an extensive range of applications and investigations in
hundreds of milliliters needed in LLE. This technique is not exhaus- analytical chemistry [13,14].
tive and only a small fraction of analytes are preconcentrated for Recently, Yangcheng and coworkers have developed a new sam-
the analysis. pling method termed directly suspended droplet microextraction
The general idea behind these novel techniques is a great (DSDME) [15]. In this method, a stirring bar is placed at the bottom
reduction in the volumetric phase ratio of the acceptor-to-donor of the aqueous sample rotating at a proper speed, which causes a
phase. This can be achieved by using either immiscible liquid weak gentle vortex or whirlpool in the solution. If a small volume of
phases (solvent microextraction) or a membrane to separate the an immiscible organic solvent is added to the surface of the solution,
acceptor–donor phases (membrane extraction). Another important the motion of the vortex results in the formation of a single micro-
advantage is the integration of extraction and injection into the drop at or near the center of rotation. The droplet itself may also
instrument in one step, thus minimizing the analysis time. Apart rotate on the surface of the aqueous phase, increasing mass transfer.
from a wide choice of extraction solvents, LPME can be performed Compared with the other LPME technique based on droplet system,
with the simplest devices, i.e. a traditional microsyringe and does i.e. single drop microextraction, it provides more flexibility in the
not suffer from carry-over between extractions that is encountered choice of the operational parameters, especially for solvent volume
when using SPME. and stirring frequency. The possibility of using larger volumes of
One of the main methodologies that evolved from the sol- organic solvent in this method in addition to GC makes it a useful
vent microextraction approach are the single drop microextraction technique comparable to HPLC and UV–vis spectrophotometry.
(SDME) technique [6–8], where the acceptor phase is a microdrop of
a water immiscible organic solvent suspended in an aqueous donor 2. Experimental
solution (two-phase system).
Another version of LPME is membrane extraction and the tech- 2.1. Reagents and standards
niques developed can be divided into two main categories: porous
membrane techniques, where the solutions on the both sides Methanol, 2-octanone and heptanol with Suprasolv quality (for
of the membrane are in physical contact through the pores of organic trace analysis) were obtained from Merck (Darmstadt, Ger-
a membrane, and non-porous membrane techniques, where the many). 1-Octanol was purchased from Fluka (Buchs, Switzerland).
membrane forms a separate phase (polymeric or liquid) between Analytical reagents grade benzene, toluene, ethylbenzene and o-
the donor and the acceptor solutions. The use of membranes xylene also were purchased from Merck (Darmstadt, Germany).
presents the advantages of high selectivity, clean extract formation To prepare stock solutions of BTEX (2000 ␮g mL−1 ) approximately
and a high degree of enrichment. One of these membrane tech- 23 ␮L of each of them was transferred into a 10-mL volumetric flask
niques was introduced by Pedersen-Bjergaard and Rasmussen [9], and dissolved with methanol. It was then stored in a refrigerator at
which was also termed hollow fiber-based liquid phase microex- 4 ◦ C. Fresh working solutions (2 ␮g mL−1 ) were prepared daily by
traction (HF-LPME). It utilized porous, hydrophobic, hollow fibers diluting the stock solution in distilled water.
impregnated with an organic phase. This new extraction methodol-
ogy proved to be an attractive alternative to other microextraction 2.2. Instrumentation
concepts because, apart from being simple, it is inexpensive, fast
and virtually solvent-free. Gas chromatographic analysis was carried out using a
SDME and HF-LPME require careful and elaborate manual opera- Chrompack CP9001 (Middelburg, the Netherlands) fitted with a
tions, given that problems of organic solvent instability/dissolution split/splitless injector and flame ionization detector (FID). Helium
have often been reported especially after faster stirring or longer (99.999%, Sabalan Co., Tehran, Iran) was used as the carrier gas
extraction time. Recently, a new mode of LPME named disper- with a flow rate of 1.11 mL min−1 . Separations were conducted
sive liquid phase microextraction (DLPME) [10], which is based using a CP-Sil 24CB (50% phenyl, 50% dimethylsiloxane) capillary
on a ternary component solvent system such as homogeneous column, WCOT Fused silica, 30 M × 0.32 mm i.d. with 0.25 ␮m sta-
liquid–liquid extraction (HLLE) and cloud-point extraction (CPE) tionary film thickness (Chrompack, Middelburg, the Netherlands).
were proposed. In this method the phenomenon of separating the The injector temperature was set at 210 ◦ C and all injections were
phase from a homogeneous solution was used and the target solutes made in the split mode. The column was initially maintained at
were extracted into a separated phase and then were determined. 60 ◦ C for 1 min; subsequently, the temperature was increased to
In DLPME, the appropriate mixture of the extraction and disperser 100 ◦ C at a rate of 5 ◦ C min−1 , then it was increased to 200 ◦ C (30 ◦ C
solvents is rapidly injected into the aqueous samples containing min−1 ). The total time for each GC run was 17 min. The FID tem-
analytes. Then, cloudy solution was formed and a drop of organic perature was maintained at 250 ◦ C. The flow of Zero Air (99.99%,
phase was sedimented in the bottom of the conical tube after cen- Sabalan Co., Tehran, Iran) for FID was 250 mL min−1 and the flow
trifugation. This sediment is withdrawn with microsyringe and rate of hydrogen was 30 mL min−1 .
introduced to an analytical instrument for further analysis.
It has been reported that the main shortcoming of the SDME, 2.3. Directly suspended droplet microextraction procedure
is the instability of the droplet when an organic solvent is used
as extractant. This fact limits the usable volume of the extracting The extraction steps are illustrated in Fig. 1. In this extraction
medium, affecting directly the precision and also the sensitivity of procedure, a 3-mL cylindrical sample cell (35 mm × 13 mm) with
the method. a screw cap, a 10-␮L syringe (Hamilton Bonaduz AG, Bonaduz,
Ionic liquids, which are ionic media resulting from the combina- Switzerland) and a 7 mm × 2 mm stir bar were used.
tion of organic cations and various anions, have been proposed as At first, 2.5 mL sample solution was held in the 3.0-mL sample
an alternative to these organic solvents due to their low vapor pres- vial, and a stirring bar was adjusted within the sample solution.
sure and their high viscosity, which allows the use of the larger and The magnetic stirrer was turned on and adjusted to a desired stir-
more reproducible extracting volumes of solvent [11]. These sol- ring speed. To make a steady and benign vortex, it is important to
vents have other unique properties, including dual natural polarity, keep the stirring bar rotating smoothly just at the center of the bot-
or miscibility with water and organic solvents [12]. Additionally, tom. A microdroplet of an immiscible organic solvent is placed at
they are regarded as environmentally friendly solvents and are the bottom of the vortex, and the syringe removed. The screw cap
938 A. Sarafraz-Yazdi et al. / Talanta 78 (2009) 936–941

Fig. 1. Photography of the different steps in DSDME: (a) magnetic stirrer is on, (b) starting organic solvent addition with a microsyringe and (c) droplet forming while stirring
bar is rotating.

should be kept closed during the extraction process. After 25 min., 3.3. Effect of the extraction time
the screw cap was removed and a portion of the organic droplet
was withdrawn into a syringe and injected into the GC for further Similar to the other LPME procedures, DSDME is a technique
analysis. which is dependent on equilibrium rather than exhaustive extrac-
tion. The amount of analyte extracted into the droplet at a given
3. Results and discussion time depends upon the mass transfer of analyte from the aque-
ous phase into the organic solvent phase. This procedure requires a
3.1. Optimization of directly suspended droplet microextraction period of time for the equilibrium to be established. However, it is
parameters not normally practical to use extraction times that are long enough
for equilibrium to be established. Fig. 3 shows the effect of extrac-
Factors affecting the extraction efficiency such as organic sol- tion time on the method efficiency. By increasing the extraction
vent, the extraction time, microdroplet volume, stirring speed and time the numbers of the moles extracted are increased, therefore
salt effect were optimized. The optimization was carried out on the peak area related to the analytes are being increased up to the
water solution of 2 ␮g mL−1 for each BTEX compounds. The chro- period of 25 min, and then decreased with the increasing of the
matographic peak area, which is related to the number of moles extraction time. This may be due to the organic solvent evaporation
of analytes which are extracted into the droplet, was used to and dissolution in water solution. Since the extraction here is not
evaluate the extraction efficiency under different experimental an exhaustive one, a reasonable period of time (25 min) is selected
conditions. Throughout these experiments, the injected volume of for the subsequent experiments.
the extracted analytes into GC was kept constant at 1 ␮L.
3.4. Microdrop volume
3.2. Choice of organic solvent
The volume of the extractor organic droplet has a great effect on
To establish a direct mode LPME technique, it is necessary to the extraction efficiency. The typical injection volume is 5–10 ␮L
choose a proper organic solvent. The choice of the organic solvent for HPLC and 50 ␮L or more for UV–vis spectrometer. Both of these
needs the following considerations: the solvent should have good volumes are beyond the upper limit of all other droplet microex-
affinity for target compounds, low solubility in water such as to pre- traction methods reported. DSDME based on free droplets and
vent the dissolution in the aqueous phase and lower density than controlled fluid fields do not fail even when using larger volumes of
water. On the basis of these considerations 1-octanol, 2-octanone organic solvent, so DSDME can well match with HPLC and UV–vis
and 1-heptanol were tested in the preliminary experiments. The spectrometer directly [15].
peak area was selected as the extraction efficiency for each solvent. The effects of the 2-octanone drop size on the extraction were
It can be seen from Fig. 2 that 2-octanone gives the best extrac- examined in the range of 7.5–15 ␮L. The relationship between the
tion efficiency and is used as the extraction solvent for subsequent volume of organic solvent and extraction efficiency are shown in
extractions. Fig. 4. Based on this trend, the analytical signals were decreased.

Fig. 2. Effect of extraction solvent on DSDME extraction efficiency (n = 3). Other Fig. 3. The effect of extraction time on the extraction efficiency of BTEX compounds
experimental conditions are as follows: concentration level at 2 ␮g mL−1 , 800 rpm when using DSDME technique with 2-octanone as solvent. Other extraction con-
stirring speed, 20 min extraction time, 2.5 mL sample volume, microdroplet volume; ditions: analyte concentration; 2 ␮g mL−1 , stirring speed; 800 rpm, 2.5 mL sample
10 ␮L. volume, microdroplet volume; 10 ␮L.
 A. Sarafraz-Yazdi et al. / Talanta 78 (2009) 936–941 939

Fig. 5. The effect of salt on the extraction efficiency of BTEX compounds. Extraction
conditions: analyte concentration; 2 ␮g mL−1 , 2-octanone as organic solvent, stirring
rate; 800 rpm, extraction time; 25 min, 2.5 mL sample volume, microdroplet volume;
7.5 ␮L.
Fig. 4. The effect of microdroplet volume on the extraction efficiency of BTEX
compounds when using DSDME technique. Extraction conditions: analyte concen-
tration; 0.2 ␮g mL−1 , 2-octanone as organic solvent, extraction time; 25 min, stirring
speed; 800 rpm, 2.5 mL sample volume.

In a smaller droplet, the area to volume ratio is greater than in a

larger droplet. Therefore, mass transfer could take place more easily
in a droplet with a smaller size. This behavior is also expected by
considering the enrichment factor equation:

CO kt
E= aq =
C0 1 + kt (V O /V aq )

aq
where CO and C0 are the analyte concentration in organic phase
and its initial concentration in the aqueous phase, respectively.
VO /Vaq is the volume ratio of organic phase to the aqueous one and Fig. 6. The effect of stirring speed on the extraction efficiency of BTEX compounds.
kt is the distribution coefficient (i.e. CO /Caq ) at time t. As can be seen Extraction conditions: analyte concentration 2 ␮g mL−1 , 2-octanone as organic sol-
from this equation, enrichment factor has a reverse correlation with vent, extraction time 25 min; NaCl concentration 0% (w/v), 2.5 mL sample volume,
microdroplet volume 7.5 ␮L.
VO /Vaq ratio. Thus 7.5 ␮L microdrop volume was chosen for further
work.
3.6. Stirring speed

3.5. Salting effects The agitation of the sample solution enhances the microextrac-
tion efficiency. In DSDME, the stirring speed has a direct influence
The extraction efficiency is related to the ionic strength of the on both the shape of the droplet and the mass transfer character-
aqueous solution [16]. Usually, depending on the solubility of the istics in the aqueous sample [15]. In Fig. 6, it is shown that the
target analytes, adding salt to the sample enhances the extraction peak areas of all analytes increase with increasing stirring speed
of the more polar analytes. In the case of DSDME, salt addition was up to 800 rpm. It was also observed that the stirring speed above
generally limiting the extraction of analytes. It was assumed that 800 rpm causes the instability and faster dissolution of the solvent
apart from the salting-out effect, salt addition causes a second effect droplet and also decreases the peak area. Hence, the stirring speed
named, salting-in effect. This phenomenon leads to changes in the of 800 rpm was chosen as the optimum stirring rate.
physical properties of the Nernst diffusion film. So, target analyte
diffusion rate into the droplet was reduced [17]. In the present work, 4. Figures of merit of the method
the effect of NaCl concentration (ranging from 0 to 15%) was inves-
tigated and the extraction efficiencies were monitored. The peak The calibration graphs were drawn using seven spiking levels of
area decreased with increasing salt concentration in the aqueous all analytes in the concentration range of 0.01–20 ␮g mL−1 . For each
sample (Fig. 5). Therefore, no salt was added to the sample solution point three replicate extractions were performed. The extraction
in the subsequent extractions. conditions were as follows: sample solution: 2.5 mL, organic sol-

Table 1
DSDME performance and validation data.

Compound Enrichment factor R.S.D.a (%) (n = 5) Linear range (␮g mL−1 ) Correlation Coefficient (r2 ) LODb (ng mL−1 )

Benzene 142.68 2.21 0.01–20 0.9996 7

Toluene 254.88 2.47 0.01–20 0.9989 3
Ethylbenzene 275.44 1.81 0.01–20 0.9991 1
o-Xylene 312.13 2.45 0.01–20 0.9995 0.8
a
Water samples spiked at 0.2 ␮g mL−1 for each compound.
b
Limit of detection for S/N = 3 (n = 5).
940 A. Sarafraz-Yazdi et al. / Talanta 78 (2009) 936–941

Table 2
Relative recoveries and precisions of DSDME technique for river water samples
spiked with the analytes.

Analytes River water Tap water

Relative recovery% R.S.D.a (%) Relative recovery% R.S.D.a (%)

Benzene 98.62 7.03 96.52 5.3

Toluene 92.65 6.5 89.35 4.78
Ethylbenzene 92.52 6.35 88.27 4.5
o-Xylene 89.87 7.3 88.78 5.68
a
Precision expressed as R.S.D. at 0.2 ␮g mL−1 level, n = 5 replicates.

Table 3
Comparison of the DSDME-GC-FID method with other related methods for determi-
nation of BTEX.

Methods LODa R.S.D. (%) Reference

HF-LPME-GC-FID 7.0–30.0 2.02–4.61 [18]

SPME-GC-FID 0.2–1.0 4–8 [19]
HS–SPME–GC–FID 0.08–0.6 3–7 [19]
DI-SDME-GC-FID 7.2–16.5 10.6–12.2 [20]
Fiber-in-tube LPME-GC-FID 0.3–5.0 3.6–8.1 [20]
DSDME-GC-FID 0.8–7.0 1.81–2.45 This study
a
LOD, limit of detection in ng mL−1 .

HS-SPME-GC-FID, DI-SDME-GC-FID, fiber-in-tube LPME-GC-FID

and HF-LPME-GC-FID, which are shown in Table 3.
Fig. 7. Chromatogram of BTEX compounds in river water spiked with 0.2 ␮g mL−1 Clearly, the relative standard deviations of DSDME-GC-FID were
of each analyte: (1) benzene, (2) toluene, (3) ethylbenzene and (4) o-xylene. better than the relative standard deviations of the other meth-
ods. The limit of detection values in DSDME-GC-FID is comparable
with the other methods. Also this method eliminates the main
vent: 2-octanone, extraction time: 25 min, stirring speed: 800 rpm, common problems encountered with SPME, such as carry-over
microdroplet volume: 7.5 ␮L. effects between analyses, limited lifetime and fragile nature of the
The calculated calibration curves gave a high level of linear- fiber.
ity for all target analytes with correlation coefficients (r2 ) ranging
between 0.9989 and 0.9996 as shown in Table 1. The repeatability 7. Conclusion
of the proposed method, expressed as relative standard deviation
(R.S.D.%), was evaluated by extracting five consecutive aqueous This technique was successfully used for the separation and
samples (spiked at 0.2 ␮g mL−1 with each target analyte) and was preconcentration of BTEX from water samples prior to analysis by
found to vary between 1.81 and 2.47%. The limits of detections capillary GC-FID. As compared with the other conventional sample-
(LODs), based on a signal-to noise ratio (S/N) of 3, ranged from 0.8 preparation methods, the DSDME-GC-FID offered advantages such
to 7 ng mL−1 (Table 1). as simplicity, ease of operation, reproducibility and relatively low
detection limit. As a conclusion, the proposed method possesses
5. Real water analysis great potential in the analysis of ultra-trace compounds in real
aqueous samples.
The DSDME technique was applied for determination of BTEX
in river water. These samples were collected from the local river of Acknowledgement
Mashhad, Iran. No BTEX compounds were detected in river water
samples; therefore, these samples were spiked with the BTEX com- The authors are grateful to Ferdowsi university of Mashhad, Iran
pounds to assess matrix effects (Fig. 7). As it is mentioned above for financial support.
“salting-in effect” causes a decrease in the extraction efficiency,
therefore for more complicate matrices the salt interferences must
References
be removed before the extraction or in the critical cases the inter-
nal standards or standard addition methods should be used. The [1] R.J. Irwin, M. VanMouwerik, L. Stevens, M.D. Seese, W. Basham, Environmental
DSDME technique is a non-exhaustive extraction procedure and Contaminants Encyclopedia, Fort Collins, Colorado, USA, 1997.
the relative recovery (determined as the ratio of the concentra- [2] R.M. Alberici, G.C. Zampronio, R.J. Poppi, M.N. Eberlin, Analyst 127 (2002)
303.
tions found in natural and distilled water samples, spiked with the [3] C. Arthur, J. Pawliszyn, Anal. Chem. 62 (1990) 2145.
same amount of analytes) instead of the absolute recovery (used [4] J. Pörschmann, F.D. Kopinke, J. Pawliszyn, J. Chromatogr. A 816 (1998) 159.
in exhaustive extraction procedures) was employed. Table 2 shows [5] H. Prosen, L. Zupančič-Kralj, Trends Anal. Chem. 18 (1999) 272.
[6] H. Liu, P.K. Dasgupta, Anal. Chem. 68 (1996 1817).
that the relative recovery and relative standard deviation (R.S.D.%)
[7] M.A. Jeannot, F.F. Cantwell, Anal. Chem. 68 (1996) 2236.
for BTEX compounds in spiked river water sample were between [8] M.A. Jeannot, F.F. Cantwell, Anal. Chem. 69 (1997) 235.
89.87 and 98.62%. [9] S. Pedersen-Bjergaard, K.E. Rasmussen, Anal. Chem. 71 (1999) 2650.
[10] M. Rezaee, Y. Assadi, M.R. Millani, E. Aghaee, F. Ahmadi, S. Berijani, J. Chromatogr.
A 1116 (2006) 1.
6. Comparison of DSDME with other related methods [11] J.-F. Liu, G.-B. Jiang, Y.-G. Chi, Y.-Q. Cai, Q.-X. Zhou, J.-T. Hu, Anal. Chem. 75 (2003)
5870.
[12] T. Welton, Chem. Rev. 99 (1999 2071).
Some statistical data of this work for water samples were com- [13] E. Aguilera-Herrador, R. Lucena, S. Cardenas, M. Valcarcel, Anal. Chem. 80 (2008)
pared with the other related methods such as: SPME-GC-FID, 793.
 A. Sarafraz-Yazdi et al. / Talanta 78 (2009) 936–941 941

[14] E. Aguilera-Herrador, R. Lucena, S. Cardenas, M. Valcarcel, J. Chromatogr. A 1201 [17] E. Psillakis, N. Kalogerakis, Trends Anal. Chem. 21 (2002) 53.
(2008) 106. [18] A. Sarafraz-Yazdi, A.H. Amiri, Z. Es’haghi, Chemosphere 71 (2008) 671.
[15] L. Yangcheng, L. Quan, L. Guangsheng, D. Youyuan, Anal. Chim. Acta 566 (2006) [19] J.C. Flórez Menéndez, M.L. Fernández Sánchez, J.E. Sánchez Uría, E. Fernández
259. Martínez, A. Sanz-Medel, Anal. Chim. Acta 415 (2000) 9.
[16] R. Eisert, J. Pawliszyn, Crit. Rev. Anal. Chem. 27 (1997) 103. [20] J.-X. Wang, D.-Q. Jiang, X.-P. Yan, Talanta 68 (2006) 945.