Analytica Chimica Acta 415 (2000) 9–20

Static headspace, solid-phase microextraction and headspace solid-phase

microextraction for BTEX determination in aqueous samples

by gas chromatography

J.C. Flórez Menéndez

a

, M.L. Fernández Sánchez

a

, J.E. Sánchez Ur´ıa

a

,

E. Fernández Mart´ınez

b

, A. Sanz-Medel

a

,∗

a

Department of Physical and Analytical Chemistry, Faculty of Chemistry, University of Oviedo, Julian Claver´ıa 8, E-33006 Oviedo, Spain

b

COGERSA, Consorcio para la Gestión de Residuos Sólidos de Asturias, La Zoreda, Ser´ın, 33697 Gijón, Spain

Received 5 January 2000; received in revised form 8 March 2000; accepted 9 March 2000

Abstract

The determination of benzene, toluene, ethylbenzene and xylenes (BTEX) in aqueous samples by gas chromatography with

flame ionization detection (GC–FID) with three sampling techniques, viz. static headspace (HS), solid-phase microextraction
(SPME) and combined HS–SPME by using a polydimethylsiloxane fiber has been investigated. Experimental parameters
optimisation of the BTEX determinations using the three techniques are given. At their optimum conditions of operation,
detection limits, times required for analysis and precision observed for the three techniques used are established. A detailed
comparison of the analytical performance characteristics of HS, SPME and HS–SPME as sampling techniques for final
GC–FID determination of BTEX in waters is given. Although all three techniques turned out to be sufficiently accurate,
offering detection limits on the

≤ng ml

−1

range, the most sensitive, selective and least time consuming technique was

HS–SPME, which seems to be particularly appropriate for routine analysis of BTEX in waters by GC. © 2000 Elsevier
Science B.V. All rights reserved.

Keywords: BTEX; Headspace; Solid-phase microextraction; Gas chromatography

1. Introduction

Volatile organic compound (VOC) determina-

tion in aqueous samples is usually carried out by
gas chromatography with flame ionisation detection
(GC–FID). The extraction step is often performed
by liquid–liquid extraction with an adequate organic
solvent. As micro-components are extracted, so the

∗

Corresponding author. Tel.:

[phone];

fax:

[phone].

E-mail address: [e-mail] (A. Sanz-Medel)

organic solvents must be of high purity, with a cor-
responding higher price. Also, many are toxic and
perhaps carcinogenic. The disadvantages of conven-
tional liquid–liquid extraction techniques are today
well documented and have led to the development of
alternative methods of trying to reduce the volumes
of organic solvents needed.

Static headspace (HS) sampling is a suitable sam-

pling technique for determining VOCs by GC. In HS,
the sample is normally placed in a sealed vial and
heated in an oven until the volatile compounds reach
equilibrium with the gas phase. The relative concen-

0003-2670/00/$ – see front matter © 2000 Elsevier Science B.V. All rights reserved.
PII: S 0 0 0 3 - 2 6 7 0 ( 0 0 ) 0 0 8 6 2 - X

10

J.C.F. Men´endez et al. / Analytica Chimica Acta 415 (2000) 9–20

trations of a given analyte in the two phases are deter-
mined by the partition coefficient, defined as the ratio
of the analyte concentration in the liquid phase to that
in the gas phase. An aliquot of the gas phase is finally
analysed by GC [1]. HS offers several advantages: it
is not expensive, does not require complicated instru-
mentation, the use of organic solvents is not necessary
and the sensitivity for benzene, toluene, ethylbenzene
and xylenes (BTEX) is dramatically enhanced com-
pared with direct injection techniques. Furthermore,
headspace analysis for BTEX and other VOCs has
been adopted by the EPA in several protocols [2,3].
However, the equipment required is not cheap and
sensitivity may not be enough for samples with very
low levels of BTEX. In fact, headspace sampling is
suitable for the analysis of samples with rather high
contents of volatile compounds (e.g. waste waters
[4–7]).

A recent advance in organic compounds analysis

is the use of solid-phase microextraction (SPME). In
this solvent-free technique, developed by Arthur and
Pawliszyn [8], analytes are adsorbed directly from
the sample onto a fused-silica fiber coated with a
polymeric phase; then the fiber is removed from the
sample solution and the analytes are thermally des-
orbed in the injector of a gas chromatograph. Hence,
sampling, extraction and preconcentration are accom-
plished in a single step. This technique is very sim-
ple, fast, portable and inexpensive [9,10]. SPME has
been applied to the determination of different pollu-
tants, such as BTEX [11–14], benzene derivatives [15],
chlorinated hydrocarbons [16,17], PAHs and PCBs
[18–21], pesticides [22,23], phenols [24,25] and inor-
ganic and organometallic compounds [26–29].

Direct sampling, that is placing the fiber directly

into the sample to extract organic compounds is rec-
ommended for relatively clean water samples. How-
ever, for dirty samples such as wastewater or aqueous
samples containing grease and/or oil, direct SPME is
not recommended, sampling of the analytes from the
headspace above the sample matrix (HS–SPME) [30]
being more adequate. So, many interference problems
are eliminated, because the fiber is not in contact with
the sample. HS–SPME is especially useful in the case
of highly volatile analytes, while for semivolatile com-
pounds, the low volatility and relatively large molecu-
lar weight may slow the mass transfer from the matrix
to the headspace, resulting in a long extraction time.

BTEX have many chemical applications and are

widely used in the manufacture of paints, synthetic
rubber, agricultural chemicals and chemical interme-
diates. BTEX compounds frequently enter soil, sedi-
ments and groundwater because of accidental oil spill,
leakage of gasoline and other petroleum fuels from
underground storage tanks and pipelines, and im-
proper oil-related waste disposal practices. BTEX are
hazardous carcinogenic and neurotoxic compounds,
and are classified as priority pollutants by the US
Environmental Protection Agency (EPA), specially
benzene, which is a leukaemic agent in humans and
has a very low tolerance standard in drinking water
of 5

␮g l

−1

[31].

A comparative study of BTEX determination by HS,

SPME and HS–SPME as sampling techniques before
final GC detection has been carried out here. Results
observed using the three techniques are compared. A
final procedure is proposed which has been applied to
the determination of BTEX in leaching waters from
tar residues.

2. Experimental

2.1. Instrumentation

Static headspace analysis was performed using a

Hewlett-Packard 7694E headspace sampler. Experi-
mental optimum parameters of the headspace sampler
were as follows: equilibration time, 25 min; extrac-
tion temperature, 70

◦

C; sample loop volume, 1 ml;

loop/transfer line temperature, 110

◦

C; sample vial

pressure, 16 psi; loop fill time, 0.03 min; injection,
1 min and sample volume, 11 ml in 22 ml vials. No
NaCl was added to the samples.

The SPME device and the fibers were obtained

from Supelco (Bellefonte, PA). The microextraction
fiber was 10 mm long, 100

␮m in diameter, and coated

with a 100

␮m thickness of polydimethylsiloxane

(PDMS). When new, the fiber was conditioned for 1 h
in a GC injector port at 250

◦

C before analysis. The

conditioned fibers were used immediately or protected
from contamination by inserting the SPME syringe
needle into a GC septum injection port before use.
For direct SPME sampling, optimum conditions were
as follows (unless otherwise indicated): sampling
time, 10 min with constant stirring to speed up phase
equilibration; room sampling temperature; desorption

J.C.F. Men´endez et al. / Analytica Chimica Acta 415 (2000) 9–20

11

temperature, 150

◦

C; desorption time, 3 min and sam-

ple volume, 10 ml in 12 ml vials. No NaCl was added
to the samples.

For HS–SPME sampling, fibers of similar character-

istics were employed. Optimum conditions were: ex-
traction time, 4 min with constant stirring to improve
the liberation of the analytes from the water matrix
to the headspace above; room sampling temperature;
sample volume, 11 ml in 22 ml vials; desorption pa-
rameters were the same as for direct SPME. Samples
were saturated with NaCl.

A Hewlett-Packard Model 5890 Series II gas chro-

matograph equipped with a flame ionization detection
(FID) system was used for the final determinations.

The GC separation column was a 50 m

×0.53 mm

i.d. DB-5 (equivalent to a 5% phenyl, 95% methylpoly-
siloxane) fused silica capillary column (J&W Sci-
entific, Folsom, CA). The carrier gas was helium at
2.3 ml min

−1

; injector and detector port temperatures

were, respectively, 150 and 320

◦

C; oven temperature

45

◦

C (hold 4.5 min), rate 30

◦

C min

−1

to 80

◦

C (hold

5 min), rate 60

◦

C min

−1

to 150

◦

C (hold 7 min).

2.2. Standards and reagents

Pure benzene, toluene, ethylbenzene, p-xylene and

o-xylene were all from Aldrich (Steinheim, Germany).

Fig. 1. Sample heating time profile for BTEX by headspace analysis (sampling temperature, 70

◦

C): benzene (

䉫

); toluene (

䊐

); ethylbenzene

(

4); m/p-xylene (×); o-xylene (*).

A primary mixture standard solution with a concentra-
tion of 3500

␮g ml

−1

of each component of BTEX was

prepared from the pure analytes by weight in methanol
(from Merck, Darmstadt, Germany). Secondary mix-
ture standard solutions were prepared, also by weight,
in ultrapure water (Milli-Q system, Millipore, Beld-
ford, MA) from the primary standard solution to give
concentrations of 7

␮g ml

−1

of each BTEX compo-

nent. Finally, optimization solutions were prepared by
weight from different dilutions of the secondary stan-
dard mixture, to give concentrations between 7 and
700 ng ml

−1

.

3. Results and discussion

3.1. Headspace analysis

3.1.1. Heating time effect

The heating-time profile for the BTEX mixture is

shown in Fig. 1. The time required for reaching equi-
librium of analytes between the liquid and gas phases
(at 70

◦

C) was 25 min, so it was chosen for subsequent

HS experiments.

3.1.2. Extraction temperature

Changes of extraction efficiency with temperature

in the range 50–90

◦

C for a 25 min extraction time

12

J.C.F. Men´endez et al. / Analytica Chimica Acta 415 (2000) 9–20

Fig. 2. Effect of loop fill time on the determination of BTEX by headspace (heating time, 25 min, extraction temperature, 70

◦

C); symbols

as in Fig. 1.

were small. In headspace analysis, it is recommended
not to use high temperatures (in order to avoid the
over-pressurisation of the vial sample, and so avoid
accidents) and, therefore, an extraction temperature of
70

◦

C was selected.

3.1.3. Effect of loop fill time

The loop fill time effect observed is shown in Fig.

2. Short times increase the signal for headspace BTEX
analysis because short times avoid equilibration of the
loop at atmospheric pressure; as the sample inside the
loop is kept at a high pressure, the analytes are more
concentrated with the subsequent improvement of the
sensitivity of the determinations. A loop fill time of
0.03 min was selected.

3.1.4. Ionic strength influence

The effect of ionic strength on the headspace BTEX

determinations was determined by saturating a sample
with NaCl. Results obtained without NaCl and when
samples were saturated in this salt were compared.
Saturation of the sample with NaCl led to just slightly
better results in the case of benzene, while for the other
BTEX no favourable, or sometimes unfavourable ef-
fects were observed. Consequently, no salt was added
for subsequent experiments.

3.1.5. Detection limits, precision, and linearity

Detection limits can be defined as the concentra-

tion of analyte in a sample at which the response
for that analyte is three times larger than the base-
line noise. The detection limits for BTEX obtained
in our experiments are shown in Table 1. The rel-
ative standard deviation (R.S.D.) ranged from 4 to
5% and good calibration linearity for all the analytes
was observed (from 10 times the detection limit to
8

␮g ml

−1

).

3.2. Solid-phase microextraction analysis

3.2.1. Stirring of the sample

Since SPME is based on an equilibrium distribution

process, the maximum amount of analytes will be ex-
tracted at the equilibrium time. Sample stirring reduces
the time required to reach equilibrium by enhancing
the diffusion of the analytes towards the fiber. Thus,
two aqueous samples with the same BTEX concen-
trations were extracted, one of them without stirring
and the other one with magnetic stirring at 900 rpm.
As shown in Fig. 3, for a 12 min extraction time,
stirring of the sample improved greatly all analyte
signals (as expected), so stirring at 900 rpm was used
throughout.

J.C.F. Men´endez et al. / Analytica Chimica Acta 415 (2000) 9–20

13

Table 1
Analytical performance characteristics of HS–GC–FID for BTEX
determinations in aqueous samples

Compound

Detection limit

Precision

a

(ng ml

−1

)

(% R.S.D.)

Benzene

1

4

b

Toluene

1

5

b

Ethylbenzene

2

4

c

m/p-Xylene

2

4

c

o-Xylene

2

4

c

a

Linear calibration range 10–8000 ng ml

−1

.

b

Five replicates at 25 ng ml

−1

level.

c

Five replicates at 50 ng ml

−1

level.

3.2.2. Extraction (absorption)-time profile

The BTEX absorption-time profile on the fiber was

tested. Results showed (see Fig. 4) that equilibrium is
reached in 3 min for benzene, 7 min for toluene and
8 min for ethylbenzene and the xylenes. Therefore,
10 min extraction was selected.

3.2.3. Extraction temperature

The extraction temperature has two opposing ef-

fects on the SPME process. Firstly, an increase in
temperature during the extraction process enhances
the diffusion of analytes towards the fiber, decreasing
the time needed to reach the equilibrium. However,
the distribution constants of the analytes are decreased

Fig. 3. Effect of the stirring speed on BTEX determinations by SPME (extraction time, 12 min); symbols as in Fig. 1.

with increasing temperature which also favours BTEX
evaporation; in other words, their concentration in
the liquid sample decreases. As shown in Fig. 5, a
temperature increase produced a negative effect on
benzene and toluene extraction (the most volatile
BTEX), not observed for ethylbenzene and xylenes.
Subsequent experiments were performed at room
temperature.

3.2.4. Ionic strength influence

It has been reported for several analytes that en-

hancements of their absorption on the fiber can be
achieved by increasing the ionic strength of the aque-
ous sample solutions by the addition of a salt (e.g.
NaCl). The effect of ionic strength on the determina-
tion of BTEX by SPME was checked by comparison
of signals when the sample was saturated with NaCl
and that obtained without adding salt. Results obtained
demonstrated that NaCl saturation of the samples pro-
duced better results only for benzene, while for the
other BTEX results were always worse. Consequently,
no salt was added to the samples for subsequent ex-
periments.

3.2.5. Desorption temperature and desorption time

The desorption-temperature profiles are shown

in Fig. 6. For benzene and toluene, a desorption

14

J.C.F. Men´endez et al. / Analytica Chimica Acta 415 (2000) 9–20

Fig. 4. Absorption-time profile for BTEX in a polydimethylsiloxane SPME fiber; symbols as in Fig. 1.

temperature of 100

◦

C is enough for total liberation

from the fiber, while for ethylbenzene and xylenes,
a temperature of 125

◦

C is required. A 150

◦

C des-

orption temperature was chosen in order to eliminate
memory effects. The desorption-time experiments
demonstrated that the time required to desorb the

Fig. 5. Effect of temperature in the absorption of BTEX into a polydimethylsiloxane SPME fiber; symbols as in Fig. 1.

analytes from the PDMS fiber is

<30 s for all the

BTEX components (constant signals were observed
for the analytes under study from 30 s to 8 min).
Nevertheless, a time of 3 min was chosen for real
samples, in order to achieve better cleaning of the
fibers for highly contaminated samples.

J.C.F. Men´endez et al. / Analytica Chimica Acta 415 (2000) 9–20

15

Fig. 6. Effect of desorption temperature on the signal area for BTEX determination by SPME; symbols as in Fig. 1.

3.2.6. Fiber position in the injector

For certain compounds, the fiber position in the GC

injector in the desorption step might be important,
probably because the injector is not uniformly heated
[9]. However, no influence was observed for benzene,
although for the rest of the BTEX a maximum was
reached at a desorption deep of 3.5 cm, which was
used for subsequent experiences.

3.2.7. Detection limits, precision, and linearity

The detection limits (3

σ criterium) attained at

optimum conditions and the precision of determi-
nations for the BTEX components are shown in
Table 2. Except for benzene, the values observed

Table 2
Analytical performance characteristics of SPME–GC–FID for
BTEX determinations in aqueous samples

Compound

Detection

Precision

Linear

limit

(% R.S.D.)

range

(ng ml

−1

)

(ng ml

−1

)

Benzene

1

8

a

10–2000

Toluene

0.3

5

b

3–2000

Ethylbenzene

0.2

4

b

2–2000

m/p-Xylene

0.2

4

b

2–2000

o-Xylene

0.2

4

b

2–2000

a

Five replicates at 15 ng ml

−1

level.

b

Five replicates at 5 ng ml

−1

level.

are clearly better than those observed for HS with
the same GC–FID determination while precision
was of the same order (around 5%). Linearity was
good, from 10 times the detection limit to 2

␮g ml

−1

;

up to 2

␮g ml

−1

memory effects on the fiber was

observed, therefore, no higher concentrations were
tested.

3.3. Headspace solid-phase microextraction
(HS–SPME)

For dirty samples the selectivity of the extraction of

BTEX by SPME can be significantly improved if sam-
pling is carried out in the headspace. In this case, only
the volatile compounds with affinity for the fiber are
adsorbed. The main HS–SPME parameters have been
identified and optimized [32] and will be investigated
here for the present analytes.

3.3.1. Stirring of the sample

In HS–SPME, the transport of analytes from the

liquid sample to the gas phase can be improved by stir-
ring the sample solution. Thus, two series of samples,
with the same concentration of BTEX, were analysed
by HS–SPME: the first one was stirred at 900 rpm,
while the second one was unstirred (extraction time
10 min). Results are plotted in Fig. 7. As expected,
magnetic stirring of the samples at 900 rpm improved

16

J.C.F. Men´endez et al. / Analytica Chimica Acta 415 (2000) 9–20

Fig. 7. Effect of the speed of stirring in BTEX determination by HS–SPME (Extraction time, 10 min); symbols as in Fig. 1.

considerably the performance of the extraction and
was used throughout.

3.3.2. Absorption-time profile

The extraction-time profiles for the BTEX compo-

nents are shown in Fig. 8. The equilibrium seems to be
reached in around 3 min for all the components and a
4 min extraction time was finally selected. This equi-
librium time is significantly shorter than that required
for SPME (as expected, because of the larger diffu-
sion coefficients of the analytes in gas phase than in
aqueous phase).

Fig. 8. Absorption-time profile for BTEX into a polydimethylsiloxane HS–SPME fiber; symbols as in Fig. 1.

3.3.3. Extraction temperature

Because partition coefficients are temperature de-

pendent, there is usually an optimum temperature for
HS–SPME. As temperature rises, more analytes are
released from the aqueous matrix to the headspace,
a process that results in higher analyte concentra-
tions in the headspace, thus favouring the HS–SPME
final determination. However, at high temperature
coating-headspace partition coefficients decrease be-
cause the absorption step is an exothermic process.
Thus, negative results can be achieved by increas-
ing the temperature. As can be seen in Fig. 9, for

J.C.F. Men´endez et al. / Analytica Chimica Acta 415 (2000) 9–20

17

Fig. 9. Effect of temperature in the absorption of BTEX into a polydimethylsiloxane SPME fiber; symbols as in Fig. 1.

BTEX better figures were achieved at low extraction
temperatures (4

◦

C). Nevertheless, in order to avoid

the use of complex thermostated baths, subsequent
experiments were carried out at room temperature
(18

±5

◦

C), which does not seem to much worsen the

recoveries as compared to 4

◦

C.

3.3.4. Ionic strength

The effect of sample ionic strength on the ab-

sorption of BTEX by a PDMS fiber for their
HS–SPME–GC–FID determination was tested by sat-
urating an aqueous sample with NaCl and comparing
the results with those obtained without salt addition.

Fig. 10. Effect of ionic strength on the determination of BTEX in waters by SPME–GC–FID.

Fig. 10 resumes the results observed. It is clear that
saturation of the sample with NaCl leads to better
extractions for all the BTEX, and so it was adopted
in subsequent experiments.

3.3.5. Desorption time, desorption temperature and
fiber position in the injector

For HS–SPME, the parameters which influence the

desorption step are the same than those mentioned
above for SPME. Therefore, a new optimisation is un-
necessary. In other words, the desorption time, temper-
ature and fiber position in the injector selected were,
respectively, 3 min, 150

◦

C and 3.5 cm, as before.

18

J.C.F. Men´endez et al. / Analytica Chimica Acta 415 (2000) 9–20

Table 3
Analytical performance characteristics of combined HS–SPME–GC–FID for BTEX determinations in aqueous samples

Compound

Detection limit (ng ml

−1

)

Precision (% R.S.D.)

Linear range (ng ml

−1

)

Benzene

0.6

7

a

6–2000

Toluene

0.2

5

b

2–2000

Ethylbenzene

0.08

3

c

0.8–2000

m/p-Xylene

0.08

3

c

0.8–2000

o-Xylene

0.08

4

c

0.8–2000

a

Five replicates at 15 ng ml

−1

level.

b

Five replicates at 5 ng ml

−1

level.

c

Five replicates at 2 ng ml

−1

level.

3.3.6. Detection limits, precision, and linearity

The observed detection limits and precisions for

BTEX are shown in Table 3. Linearity was good from
10 times the detection limit to 2

␮g ml

−1

; at 2

␮g ml

−1

memory effects on the fiber were observed, therefore,
no higher concentrations were tested.

3.4. Comparison of HS, SPME and HS–SPME

According to our results for BTEX determination,

the lowest detection limits are provided by HS–SPME.
Also, the time required for the analysis turned out to
be 25 min for HS, 10 min for SPME and 4 min for
HS–SPME. In other words, the last technique provides
the shortest analysis time (higher sample throughput).

Sample treatment is minimum in the three tech-

niques, in fact, no treatment at all is required for HS
and SPME, while for HS–SPME only saturation of

Fig. 11. HS–SPME chromatogram of a sample of contaminated water resulting from the leaching of tar residues (benzene (1); toluene (2);
ethylbenzene (3); m

+p-xylene (4); o-xylene and unknown compound (5)).

the samples with NaCl is recommended (not strictly
needed).

Three samples of three waters resulting from

the leaching of tar residues were analysed by the
three proposed methods using the standard additions
method. As m/p-xylene in Sample 1 is in a concentra-
tion range significantly higher than the other BTEX,
different amounts of analytes had to be added for a
correct quantification; so for benzene, toluene and
ethylbenzene the range of additions was 50, 100, 250,
500 and 750 ng ml

−1

, while for m/p-xylene the range

of additions was 250, 500, 750 and 1000 ng ml

−1

. For

samples 2 and 3, the concentrations added were, for
all compounds, 50, 100, 200, 300 and 400 ng ml

−1

.

Fig. 11 shows the type of chromatograms obtained
for such real samples by HS–SPME with GC–FID fi-
nal determination. As shown in Fig. 11, the o-xylene
peak overlapped with that of an unknown com-

J.C.F. Men´endez et al. / Analytica Chimica Acta 415 (2000) 9–20

19

Table 4
GC–FID determination of BTEX in contaminated waters resulting from of tar residues leaching with the three different sampling systems

Sample no.

Compound

Concentration (ng ml

−1

)

a

HS

SPME

HS–SPME

1

Benzene

189

±4

163

±23

179

±12

Toluene

580

±17

562

±29

569

±21

Ethylbenzene

350

±8

343

±22

361

±16

(m

+p)-Xylene

1125

±46

1018

±75

1094

±101

2

Benzene

95

±3

83

±8

94

±5

Toluene

228

±8

212

±15

236

±12

Ethylbenzene

161

±4

162

±6

173

±11

(m

+p)-Xylene

293

±3

277

±14

285

±9

3

Benzene

65

±4

56

±8

59

±6

Toluene

139

±2

127

±12

125

±9

Ethylbenzene

96

±2

88

±9

91

±6

(m

+p)-Xylene

200

±3

184

±23

188

±9

a

Mean

±S.D. (n=3).

pound, and could not be determined. The other
BTEX components were not interfered with. Results
in Table 4 show good concordance for the three
techniques.

4. Conclusions

The research presented has demonstrated that HS,

SPME and HS–SPME are solventless extraction tech-
niques alternative to more conventional ones. Extrac-
tion and preconcentration of BTEX are carried out in
a single step, and so detection limits in the range of
ng ml

−1

and pg ml

−1

are easily achieved.

The three sample pretreatment techniques under

scrutiny, namely HS, SPME and HS–SPME, seem
to provide accurate determinations for BTEX in
real samples, as inferred from the equivalent re-
sults obtained by the three techniques, two of them
(HS and SPME) with absolutely different funda-
ments. However, for the determination of BTEX in
contaminated waters, HS–SPME affords the low-
est detection limit with the shortest time required
to reach equilibrium. Therefore, it appears that this
latter combined technique can safely be recom-
mended for contaminated waters and to increase
sample throughput in routine analysis of BTEX
by GC.

Acknowledgements

The authors thank FICYT (Asturias, Spain) for fi-

nancial support under the frame of the Project Ref.
PA-AMB97-04.

References

[1] Z. Penton, J. High Resolution Chromatogr. 15 (1992) 834.
[2] A.D. Hewitt, Environ. Sci. Technol. 32 (1998) 143.
[3] T.C. Voice, Y.B. Kolb, J. Chromatogr. Sci. 32 (1994) 306.
[4] R.A. Ketola, V.T. Virkki, M. Ojala, V. Komppa, T. Kotiaho,

Talanta 44 (1997) 373.

[5] C. Zwiener, F.H. Frimmel, Fresenius J. Anal. Chem. 360

(1998) 820.

[6] B. Kolb, J. Chromatogr. A 842 (1999) 163.
[7] P. Kuran, L. Sojak, J. Chromatogr. A 733 (1996) 119.
[8] C.L. Arthur, J. Pawliszyn, Anal. Chem. 62 (1990) 2145.
[9] H. Prosen, L. Zupancic-Kralj, Trends Anal. Chem. 18 (1999)

272.

[10] A. Peñalver, E. Pocurull, F. Borrull, R.M. Marcé, Trends

Anal. Chem. 18 (1999) 557.

[11] T. Górecki, J. Pawliszyn, Anal. Chem. 67 (1995) 3265.
[12] C.L. Arthur, L.M. Killiam, K.D. Buchholz, J. Pawliszyn,

Anal. Chem. 64 (1992) 1960.

[13] C.L. Arthur, L.M. Killiam, S. Motlagh, M. Lim, D.W. Potter,

J. Pawliszyn, Environ. Sci. Technol. 26 (1992) 979.

[14] B.L. Wittkamp, D.C. Tilotta, Anal. Chem. 67 (1995) 600.
[15] S. Huang, C. Cheng, Y. Sung, Anal. Chim. Acta 343 (1997)

101.

[16] B.D. Page, G. Lacroix, J. Chromatogr. 648 (1993) 199.

20

J.C.F. Men´endez et al. / Analytica Chimica Acta 415 (2000) 9–20

[17] R. Krska, K. Taga, R. Kellner, Appl. Spectrosc. 47 (1993)

1484.

[18] B.D. Page, G. Lacroix, J. Chromatogr. A 757 (1997) 173.
[19] K. Poon, P.K.S. Lam, M.H.W. Lam, Anal. Chim. Acta 396

(1999) 303.

[20] D.W. Potter, J. Pawliszyn, Environ. Sci. Technol. 28 (1994)

298.

[21] Z. Zhang, J. Pawliszyn, J. High Resolution Chromatogr. 16

(1993) 689.

[22] A.A. Boyd-Boland, S. Magdic, J. Pawliszyn, Analyst 121

(1996) 929.

[23] S. Magdic, J. Pawliszyn, J. Chromatogr. A 723 (1996) 111.
[24] P. Barták, L. Cap, J. Chromatogr. A 767 (1997) 171.

[25] K.D. Buchholz, J. Pawliszyn, Anal. Chem. 66 (1994)

160.

[26] E. Out, J. Pawliszyn, J. Microchim. Acta 112 (1993) 41.
[27] T. Górecki, J. Pawliszyn, Anal. Chem. 68 (1996) 3008.
[28] G. Lespes, V. Desauziers, M. Potin-Gautier, J. Chromatogr.

A 826 (1998) 67.

[29] Y. Cai, J. Bayona, J. Chromatogr. 696 (1995) 113.
[30] Z. Zhang, M.J. Yang, J. Pawliszyn, Anal. Chem. 66 (1994)

844A.

[31] Z. Wang, M. Fingas, M. Landriault, L. Sigouin, N. Xu, Anal.

Chem. 67 (1995) 3494.

[32] M.E. Miller, J.D. Stuort, S.R. Smith, W. Widmer, Pittsburg

Conference, March 1997, 277 pp. (Abstract).