You are on page 1of 7

Microchemical Journal 108 (2013) 117123

Contents lists available at SciVerse ScienceDirect

Microchemical Journal
journal homepage: www.elsevier.com/locate/microc

Application and optimization of headspace solid-phase microextraction (HS-SPME)


coupled with gas chromatographyame-ionization detector (GCFID) to determine
products of the petroleum industry in aqueous samples
Maciej Tankiewicz a,, Calum Morrison b, Marek Biziuk a
a
Department of Analytical Chemistry, Faculty of Chemistry, Gdask University of Technology, ul. G. Narutowicza 11/12, 80-233 Gdask, Poland
b
School of Science, University of the West of Scotland, Paisley Campus, Paisley PA1 2BE, UK

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

Article history: A simple and solvent-free analytical method has been optimized for the determination of hydrocarbons (C6C20)
Received 19 July 2012 from products of the petroleum industry in aqueous samples. Extraction and analysis were carried out using
Received in revised form 2 October 2012 headspace solid-phase microextraction (HS-SPME) and gas chromatography coupled with ame-ionization de-
Accepted 18 October 2012
tector (GCFID). A ZB-5 capillary column (30 m0.25 mm inner diameter, 0.25 m lm thickness) was used for
Available online 26 October 2012
GC separation and a 100 m polydimethylsiloxane (PDMS) ber was used to isolate the analytes. Parameters of
Keywords:
HS-SPME, analytes desorption and GCFID were selected and an analytical procedure is proposed. Under the op-
Headspace solid-phase microextraction timum conditions the limits of detection were in the range of 2.013 g L1 and the relative standard deviations
(HS-SPME) were between 2.6 and 14%. The developed analytical method was successfully applied to the analysis of water
Gas chromatography contaminated by unleaded petrol and unltered water used to extinguish a simulated re.
Petroleum products 2012 Elsevier B.V. All rights reserved.
Hydrocarbons (C6C20)
Water samples

1. Introduction sample loss and improving detection limits. Numerous extraction


techniques are available often using liquidliquid extraction (LLE).
The continuous development of human activities means that the Commonly used solvents for extraction are dichloromethane [10],
environment is increasingly exposed to organic pollutants which nd ethers and carbon disulde [1114]. Although LLE is relatively sim-
their way into watercourses. One of the most serious threats exists ple and inexpensive, it has many drawbacks, among them the need
from products of the petroleum industry which can result in environ- to use large quantities of solvents. In order to achieve the desired
mental contamination and toxicity to life. Products of the petroleum enrichment factor, the excess solvent requires removal by evapora-
industry can also constitute a re hazard and may be used as accelerants tion, and extract cleanup may also be necessary. Additionally, there
by arsonists to increase the rate and intensity of a re [1,2]. The most is a risk of emulsions forming during agitation. This technique is useful
commonly used accelerants are liquids, particularly petrol, diesel, for extracting empty containers and for small samples [15]. Huang et al.
kerosene and complex mixtures of hydrocarbons obtained from the [16] applied a second form of solvent extraction for arson debris using su-
fractional distillation of crude oil due to their low cost and ready percritical uid extraction (SFE). The recoveries of various ignitable liquid
availability [38]. residues from carpet were above 80% with a low degree of co-extraction
Hydrocarbon products which originate from the petroleum industry of interfering compounds. Other traditionally used sample preparation
are complex in terms of chemical composition and pose analytical chal- techniques are steam and vacuum distillation, which have been applied
lenges in terms of extraction and analysis. For example accelerants may to complex matrices. Steam distillation [17] is used for large amounts of
be composed of hundreds of compounds and analysis of these products ignitable liquids in samples and commonly, the extraction solvent is
may be complicated after a re by the presence of pyrolysis products n-hexane, which provides high recovery efciency. Vacuum distillation
formed from the combustion of common household items including is useful in the case of fragile samples such as burned documents. Howev-
plastics, woods and carpets [9]. Therefore, it is necessary to prepare er, the time of distillation and use of hazardous solvents outweigh the
samples for analysis by isolating the components from a complex ma- benets [2]. In addition the pre-concentration of analytes on activated
trix and enrichment prior to nal determination. Ideally, this approach charcoal strips (ACS) [18,19] and Tenax GC [20,21] have been the most
should minimize any possible background interferences, while avoiding commonly used sorbents for evaluation, but may require the use of
hazardous solvents and/or additional thermal desorption equipment.
Corresponding author. Tel.: +48 58 347 21 10; fax: +48 58 347 26 94. Single-drop microextraction (SDME) is a simple, inexpensive meth-
E-mail address: maciej.tankiewicz@gmail.com (M. Tankiewicz). od of preparing samples with analyte isolation and preconcentration

0026-265X/$ see front matter 2012 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.microc.2012.10.010
118 M. Tankiewicz et al. / Microchemical Journal 108 (2013) 117123

taking place in a single step [22]. This technique was applied by Xiong A standard stock solution was prepared by dissolving hexane, tol-
et al. [23] to determine light hydrocarbons C6C12 in petroleum and uene, o-xylene, m-xylene, p-xylene, nonane, decane, undecane,
aqueous samples. Under optimum extraction conditions (1.5 L dodecane, tetradecane, pentadecane, hexadecane and eicosane in a
drop of n-hexadecane, 30 min extraction of 5 mL aqueous sample mixture of ethyl acetate and methanol 10:90 (v/v) at the concentra-
and stirring at 1000 rpm at room temperature) limits of detection tion 2000 mg L 1 and stored at 4 C. The standard aqueous samples
(LODs) were in the range of 0.264.33 g L 1 and relative standard were produced by spiking puried water with the standard solution
deviations (RSD) were 3.86.1%. In comparison with traditional at different concentration levels. For optimization of HS-SPME, the
LLE, the ratio of solvent to sample is small, and SDME is easily ap- working solutions were prepared by spiking puried water with
plied. However, some disadvantages have been noted and included: the standard solution at 10 mg L 1 concentration level for each
the difculty of achieving a stable organic drop, air bubble formation, hydrocarbon.
equilibrium process is slow resulting in time-consuming extraction
in most cases. 2.2. Real water samples
Solid-phase microextraction (SPME) was developed in 1989 [24]
and is a solventless extraction and which has been used to identify pe- To verify the suitability of the developed analytical procedure two
troleum products in real arson-suspected re debris samples [9,2533]. types of real water samples were used. Firstly, tap water contaminated
SPME is an excellent alternative for traditional extraction techniques, by unleaded petrol 0.1% (v/v) was used to examine the applicability of
and can be divided into Direct Immersion (DI-SPME) and Headspace the HS-SPME method for samples containing high concentrations of dif-
(HS-SPME) types, depending on where the ber is situated. Many of ferent hydrocarbons. Water from reghting was used in this study to
the advantages of SDME also apply here, for example SPME reduces calculate the recovery of analytes and check the impact of pyrolysis
the extraction time by 1020 min in comparison with activated char- products on analytical results and the SPME ber. For this purpose, a
coal strips (ACS) [34], shows greater sensitivity in detecting gasoline simulated re on a small scale was performed in the laboratory. The fol-
in a study at levels where other passive methods have failed [29] and lowing frequently occurring household materials were added to the
is unaffected by interferences from the pyrolysis products of wood re: plastic, wood, paper, pieces of carpet, glass, cotton garments and
and plastics [35]. In addition headspace sampling eliminates interfer- rubber products. To extinguish the re 2 L of tap water was used,
ences, can be used when aqueous media are present and has the ability which was then stored unltered at 4 C.
to be used in the eld. Therefore in re investigation, HS-SPME is poten-
tially useful due to complex matrix composition and potentially high 2.3. HS-SPME procedure
water content. DI-SPME has been applied to the recovery of accelerants
and explosives from aqueous samples [36,37], but a SPME method has HS-SPME extraction was performed using a Supelco (Bellefonte,
not been reported so far. Water samples represent 5% of all samples USA) manual holder containing a ber assembly with a 100 m polydi-
for re investigation [9] and liquid matrix membrane SPME has been methylsiloxane (PDMS) coating. The ber was conditioned in the injec-
applied to complex and dirty matrices [38]. Direct-contact SPME has tion port of a gas chromatograph at 250 C for 30 min prior to use and
been applied to soil by Yoshida et al. [39], with the technique also blank desorption was performed. 2 mL of aqueous sample (standard
used for identication of ignitable liquids from human skin [27]. The or real) was placed in 4 mL glass vial with PTFE/silicone septa, which
SPME method, compared with traditional LLE, limits loss of petroleum contained 10% (w/v) sodium chloride salt. The ber was exposed to
products during the sampling procedure since the process incorporates the headspace of a sample, kept at a selected temperature for a set
extraction, concentration and injection into a single step, which, in turn, time, retracted into a needle, and introduced into the injection port of
makes it more time efcient. Additionally, the HS mode provides a lon- gas chromatograph. Extraction was supported by magnetic stirring at
ger lifetime of the ber and protects the ber from any adverse effects the highest possible level. A ber blank experiment was performed to
that may occur due to non-volatile, high molecular weight substances ensure that there were no contaminants on the ber.
in the real samples. This allows determination of petroleum products For optimization of HS-SPME process, the factors affecting extrac-
in low levels of concentrations with greater accuracy and precision. tion efciency (headspace volume, extraction time, temperature,
The objective of this study was to develop and apply a procedure sample ionic strength and agitation speed) were investigated.
to determine products of the petroleum industry in aqueous samples
using headspace SPME with a ber for sampling and GCFID for anal- 2.4. Gas chromatography ame-ionization detector (GCFID)
ysis. Parameters of extraction (time, temperature, agitation rate, con-
tent of salt), and the time and temperature of desorption in the gas HP5890 series II gas chromatograph equipped with a split/splitless
chromatograph injector were selected. The procedure was validated injector coupled with a ame-ionization detector connected to an
and applied to analyze water contaminated by unleaded petrol and Agilent 3396 series 3 integrator was used. Analytes were desorbed
water used to extinguish a simulated re. from the ber in a gas chromatograph splitless injection port. The injec-
tion port was tted with a 0.75 mm i.d. injection liner (Supelco) with
2. Materials and methods the splitsplitless purge valve opened at 1 min after injection. The de-
sorption temperature and time of analytes were optimized. The depth
2.1. Reagents and chemicals of the SPME needle in the injection port was also selected (3 cm) and
controlled. The injector temperature was 260 C. The chromato-
Hexane (99%), o-xylene (98 +%), m-xylene (98 +%), p-xylene graphic column used was Zebron ZB-5 (30 m 0.25 mm 0.25 m
(98 +%), nonane (99%), decane (99 +%), undecane (99%), dodecane of 5% phenyl95% dimethylpolysiloxane) supplied by Phenomenex
(99%), pentadecane (99%), hexadecane (99%), eicosane (99%) and (USA). Helium (CP grade, 99.999%) was used as a carrier gas and was
sodium chloride were purchased from Sigma-Aldrich (Poole, UK). supplied by BOC Gases Europe (Guildford, UK). The ow of carrier gas
Toluene (99%), methanol and ethyl acetate were supplied by Fisher was 1 mL min1, with constant ow conditions being observed
Scientic (Loughborough, UK). Tetradecane (99%) was purchased throughout. The temperature program was as follows: 40 C (hold
from Koch-Light Laboratories Ltd. (Colnbrook, UK) and unleaded 2 min), 10 C min1 to 70 C, 5 C min1 to 90 C, then 25 C min1
petrol from local Shell station (Paisley, UK). All reagents, except pet- to 270 C (hold 1 min). The detector temperature was 300 C. For conr-
rol, were of analytical or higher grade and solvents were of HPLC mation of the developed analytical methodology a second fused silica
grade. The water used in this study was obtained from an ultrapure column was also used: Equity-1 capillary column (30 m0.25 mm
water purication system Elgastat, Elga Ltd. (Bucks, UK). 0.25 m lm thickness) supplied by Supelco.
M. Tankiewicz et al. / Microchemical Journal 108 (2013) 117123 119

3. Results and discussion required for the extraction process is one of the most important
parameters, which should be investigated, because HS-SPME is an
3.1. Gas chromatography (GC) separation equilibrium process of analytes between the ber coating and vapor
phase. In this study the selected time periods were 5, 15, 30, 45 and
Since hydrocarbons C6C20 are apolar compounds columns with 60 min. Each data point is the average of three independent measure-
apolar properties were selected for this study. In order to achieve sepa- ments. The areas of peaks obtained for each hydrocarbon with the
ration of all analytes and appropriate peak shapes a temperature pro- different extraction time are presented in Fig. 1.
gram was optimized. Firstly, an experiment with a constant increasing The best extraction efciency for all studied hydrocarbons was
of temperature was performed in order to obtain the retention times observed when a longer extraction time was used. Improved results
of analytes, and then introduced three different temperature ramps in were achieved up to 45 min, except toluene, for which there were
order to achieved full separation of analytes. Unfortunately, no separa- no differences between the 30 and 45 min extraction times. Based
tion of p-xylene and m-xylene was achieved. Both compounds were elut- on the experimental results, the extraction time giving the highest re-
ed from the column at the same time. This is caused by these analytes sponses was 45 min. After this time the extraction efciency de-
having a similar boiling point. However, separation from o-xylene was creased, except in the case of tetradecane, pentadecane, hexadecane
achieved successfully. Therefore, in further analysis both p-xylene and and eicosane. For these hydrocarbons slightly better efciency was
m-xylene compounds were treated as one analyte p-(m)-xylene. The observed after 60 min. This may be due to their lower diffusion coef-
ow rate of carrier gas was also optimized. Conditions of 1, 2 and cient and higher molecular mass than other hydrocarbons. The ex-
4 mL min1 were tested and a ow rate of 1 mL min1 gave the best traction process time of 45 min was therefore chosen.
results.
3.3.3. Extraction temperature
3.2. Desorption of analytes The extraction temperature affects the efciency of the extraction
process by controlling the diffusion rate of hydrocarbons into the coat-
Hydrocarbons C6C20 were isolated into the PDMS ber coated ing of the ber. The inuence of temperature on the area of peaks for
from the headspace of an aqueous standard sample of 10 mg L 1 of each analyte was investigated varying the temperature of extraction
each analyte and desorbed for 3, 5 and 7 min at injector temperatures between 23 C and 90 C with a constant extraction time of 45 min. In
ranging from 210 to 270 C (10 C increments) with measurement of this experiment room temperature, 45, 65 and 90 C were chosen for
peak areas; 5 min desorption at 260 C was sufcient to remove all investigation. The detector responses (areas) for each hydrocarbon
analytes from the ber and was selected for further analysis. Under with different extraction temperatures are presented in Fig. 2.
these conditions the desorption process from the ber was performed As can be seen in Fig. 2 the higher the extraction temperature is
twice for each extraction event. Standard samples of three different above room temperature the lower the efciency of the extraction.
concentrations of each hydrocarbon, i.e., 10, 1, and 0.1 mg L 1, The results showed that increasing the temperature causes lower de-
were analyzed. Peak areas of particular hydrocarbons on the second tector responses (areas) for all hydrocarbons, except tetradecane,
desorption chromatograms were never larger than 0.5% of those pentadecane, hexadecane and eicosane. For these analytes the
obtained in the rst desorption. This suggests that only trace, if any, highest extraction efciency was observed at 65 C. The optimum
amounts of extracted C6C20 remained on the ber after the rst de- extraction efciency was achieved at ambient temperature and was
sorption event and can be neglected as they are smaller than the ex- selected for the subsequent experiments.
perimental errors.

3.3. Optimization of HS-SPME conditions 3.3.4. Salt addition


The inuence of sodium chloride (NaCl) salt additives to the
HS-SPME is an equilibrium process that involves the partitioning of HS-SPME procedure was investigated by comparing the extraction ef-
analytes from aqueous phase to the gas phase and into the polymeric ciency of samples with different concentrations of NaCl. In this study
phase according to their partition coefcients. Thus, the optimization 5, 10 and 20% (w/v) of salt was used. Higher salt content has not been
of parameters that inuence the partition of hydrocarbons C6C20 be- taken into consideration because of problems with dissolving NaCl.
tween the headspace and the solution are very important. For Fig. 3 presents the effect of NaCl additives on detector responses
HS-SPME, the critical factors that affect extraction efciency include (areas).
headspace volume, extraction time, temperature, sample ionic strength The addition of salt into the aqueous sample prior to the extraction
and agitation speed. All of these were taken into consideration in process increased the ionic strength of the solution. As a consequence,
this study. Each data point is the average of three independent the diffusion of the analytes into the headspace is favored and extrac-
measurements. tion time is reduced. As can be seen in Fig. 3 a slight increase in peak
areas was observed with the increasing of NaCl concentration up to
3.3.1. Effect of headspace volume 10%. For higher content of NaCl the extraction efciency decreased.
Firstly the effects of the water sample and headspace volume were Thus, the optimum concentration of salt for the extraction of hydrocar-
studied. The experiment was performed using 4 mL vials and the vol- bons was 10% (w/v) and was chosen for further analysis.
ume of water was increased from 1 to 3 mL (headspace volume from
3 to 1 mL). A standard solution with the concentration of 10 mg L 1 3.3.5. Magnetic stirring
for each analyte was analyzed twice. The extraction time was 30 min Magnetic stirring accelerates extraction and shortens the time
at the ambient temperature and desorption time was 5 min at 260 C. needed to reach thermodynamic equilibrium between the aqueous
Little effect of the ratio of sample volume to headspace volume on the and the headspace phase. For headspace, stirring should be vigorous
extraction efciency was observed. The optimum detector response and constant to provide excellent reproducibility. Generally, the larger
was found to be 2 mL of spiked water sample with 2 mL of headspace the agitation, the faster the mass transfer between the aqueous phase
volume and so was used for subsequent analysis. and the headspace. The stirring speed range studied was 01000 rpm.
From the obtained results it can be stated that extraction efciency
3.3.2. Extraction time was poor without stirring and the detector responses (areas) were
An evaluation experiment was performed using different extrac- higher for each analyte when the stirring rate increased. An optimal
tion times from 5 to 60 min at the ambient temperature. The time stirring speed of 1000 rpm (the highest possible level) was used in
120 M. Tankiewicz et al. / Microchemical Journal 108 (2013) 117123

Fig. 1. Effect of extraction time from 5 to 60 min at ambient temperature on extraction efciency: desorption time 5 min at 260 C, concentration of each hydrocarbon 10 mg L1,
n = 3.

the subsequent experiments to ensure the efcient extraction of all hexadecane, and up to 1000 g L 1 for toluene, p-(m)-xylene,
hydrocarbons. o-xylene, decane, undecane, dodecane and eicosane. Coefcient of
variability (percentage of relative standard deviation, CV %) was
3.4. Method validation the average value of different concentrations of studied hydrocar-
bons in the linear range and was in the range from 0.2% (decane)
The optimized conditions were used for validating the HS-SPME to 9.4% (eicosane), which is considered as good method precision.
method for quantitative analysis of C6C20 hydrocarbons (linear range, The sensitivity of the HS-SPME method was considered in terms of
detection limit, and precision). The working standard solutions for the limit of detection (LOD). LODs were calculated from calibration func-
calibration study were prepared by spiking puried water with the stan- tions [40] using Eq. (1):
dard solution in the concentration range of 11000 g L1 for each ana-
lyte. The linear range for hydrocarbons was studied by replicate analysis 3:3 S
LOD 1
of the standard stock solutions. The linear regression values were calcu- b
lated with the average peak areas of three replicate injections for each
analyte. Table 1 presents the linear regression for each analyte with coef- where S is the residual standard deviation of the calibration function
cient of determination in the range from 0.9281 (eicosane) to 0.9983 and b is the slope of the rst linear function. As can be seen in Table 1
(nonane). the method allows detection of the selected hydrocarbons in water
The calculated calibration curves showed good linearity up samples at concentrations lower than 12.5 g L 1. LODs were in the
to 500 g L 1 for hexane, nonane, tetradecane, pentadecane and range from 2.0 g L 1 (eicosane) to 12.5 g L 1 (pentadecane). The

Fig. 2. Effect of extraction temperature from 23 C to 90 C on extraction efciency: extraction time 45 min, concentration of each hydrocarbon 10 mg L1, n= 3.
M. Tankiewicz et al. / Microchemical Journal 108 (2013) 117123 121

Fig. 3. Effect of content of sodium chloride (NaCl) salt (%) (w/v) on extraction efciency: extraction time 45 min, ambient temperature, concentration of each hydrocarbon
10 mg L1, n = 3.

limits of quantitation (LOQs) dened as 3 times the LOD were analyte- concentration level 40 g L 1 for each analyte (b) and spiked
dependent and they were in the range from 6.1 g L1 (eicosane) to water from reghting at the concentration level 200 g L 1 for
38 g L 1 (pentadecane). each analyte (c) using HS-SPMEGCFID.
Reproducibility of the method was examined by extracting ve real The conrmation of the developed methodology using an Equity
water samples from reghting spiked at two different concentration -1 capillary column (30 m 0.25 mm 0.25 m lm thickness)
levels 40 g L 1 and 200 g L 1 under optimized experimental condi- was performed, as identication based solely on retention times is in-
tions. As shown in Table 2, RSDs vary between 0.037 (hexadecane) and sufcient. In this experiment an apolar column was used. Conditions
0.14 (pentadecane) for the hydrocarbons at the 40 g L 1 concentra- and parameters during the HS-SPME process and nal determination
tion, and between 0.026 (o-xylene) and 0.10 (hexadecane) for the by GCFID were the same as outlined previously. From the obtained
hydrocarbons at the 200 g L 1 concentration, indicating that the re- results it can be stated that all hydrocarbons were sufciently sepa-
producibility of this extraction method is satisfactory. rated, except p-xylene and m-xylene which were eluted from the col-
The mean recoveries obtained for the studied hydrocarbons umn at the same time.
spiked at the two concentration levels 40 and 200 g L 1 in the real
water samples are shown in Table 2. The relative recovery, as deter- 3.5. Applications
mined by the peak area ratio of real sample to puried water sample
spiked with hydrocarbons at the same concentration level, was The developed procedure was applied to determine the hydrocar-
applied because SPME is a non-exhaustive extraction technique in- bons C6C20 in two real aqueous samples. Firstly, tap water contaminat-
stead of an absolute recovery as used in exhaustive extraction tech- ed by unleaded petrol 0.1% (v/v) was used to examine the applicability
niques. The recovery of analytes at the concentration level of of the HS-SPME method for samples containing high concentrations of
40 g L 1 ranged between 98% (pentadecane) and 118% (hexane), different hydrocarbons. Approximately 2 L of unleaded petrol was
except for decane for which the recovery was 93%. For hydrocarbons added to 4 mL glass vial. The vial was then lled with a volume of tap
C6C20 added to the water samples at the concentration level of water up to 2 mL, which contained 10% (w/v) of sodium chloride salt
200 g L 1, the recovery was in the range of 87% (undecane) to and extracted under optimized conditions. Fig. 5 shows the chromato-
117% (hexane). Fig. 4 displays the chromatograms obtained for gram obtained for the contaminated tap water sample by 0.1% (v/v) of
water from reghting (a), spiked water from reghting at the unleaded petrol using the HS-SPMEGCFID.

Table 1
Basic validation parameters obtained for each hydrocarbon by using HS-SPMEGCFID.

Analyte Retention time Equation Coefcient of Limit of detection LOD Limit of quantication Linearity range Coefcient of
[min] determination R2 [g L1] LOQ [g L1] [g L1] variability CV [%]

Hexane 2.427 y = 2791 x + 42,461 0.9959 2.4 7.3 7.3500 1.8


Toluene 3.931 y = 20,602 x 363,412 0.9864 3.7 11 111000 7.2
p-(m)-xylene 5.910 y = 51,124 x + 182,587 0.9926 6.5 19 191000 1.9
o-xylene 6.429 y = 31,568 x 655,764 0.9859 3.2 9.6 9.61000 1.7
Nonane 6.558 y = 16,891 x + 302,798 0.9983 6.9 21 21500 0.5
Decane 9.025 y = 58,668 x 2,228,190 0.9759 3.6 11 111000 0.2
Undecane 10.876 y = 52,817 x 2,071,517 0.9750 4.5 14 141000 2.5
Dodecane 12.067 y = 52,582 x 3,152,724 0.9618 7.6 23 231000 0.4
Tetradecane 13.702 y = 8305 x + 1,120,376 0.9394 7.2 22 22500 2.8
Pentadecane 14.341 y = 7810 x + 937,236 0.9654 12.5 38 38500 5.0
Hexadecane 14.917 y = 7237 x + 943,146 0.9448 5.2 16 16500 4.3
Eicosane 16.900 y = 964 x + 297,369 0.9281 2.0 6.1 6.11000 9.4
122 M. Tankiewicz et al. / Microchemical Journal 108 (2013) 117123

Table 2 and products of pyrolysis from the different kinds of materials present
Mean recovery and RSD value of selected hydrocarbons C6C20 in real water samples at a re scene did not affect for the results.
from reghting by using HS-SPMEGCFID (n = 5).

Analyte Spiked water at Spiked water at 4. Conclusion


concentration 40 g L1 concentration 200 g L1

Recovery [%] RSD (n = 5) Recovery [%] RSD (n = 5) In this study, the HS-SPME method for the extraction and GCFID
Hexane 118 0.039 117 0.041 for the nal determination of the products of the petroleum industry
Toluene 115 0.091 102 0.056 were developed. Extraction conditions were optimized to 45 min,
p-(m)-xylene 101 0.11 106 0.087 2 mL aqueous sample, 10% (w/v) of sodium chloride salt, the highest
o-xylene 107 0.078 101 0.026
possible stirring rate and room temperature. Under the optimized
Nonane 109 0.094 89 0.082
Decane 93 0.10 93 0.043
conditions, the LODs were in the range from 2.0 g L 1 (eicosane)
Undecane 110 0.058 87 0.061 to 12.5 g L 1 (pentadecane). The developed analytical method was
Dodecane 102 0.064 91 0.039 successfully applied to the analysis of water contaminated with un-
Tetradecane 99 0.041 104 0.085 leaded petrol and unltered water used to extinguish a simulated
Pentadecane 98 0.14 111 0.097
re. The results show that the selectivity of developed methodology
Hexadecane 113 0.037 108 0.10
Eicosane 103 0.12 101 0.086 is sufcient to analyze samples of very complex matrices. Pyrolysis
products from differing materials did not affect the results. Therefore,
the HS-SPME can be an effective tool for the sampling of products of
As can be seen in Fig. 5, eight hydrocarbons were detected: hex- the petroleum industry. The developed analytical method here is sim-
ane, toluene, p-(m)-xylene, o-xylene, nonane, decane, undecane and pler, lower cost and less labor intensive, than conventional tech-
dodecane. The results show that the selectivity of the developed niques such as LLE and ACS for the simultaneous determination of
methodology is sufcient to analyze samples of very complex matri- hydrocarbon analytes by GC. The lack of solvents and ability to use
ces and can be used in monitoring studies to control the content of the injection port of a gas chromatograph for thermal desorption
hydrocarbons C6C20 in environmental water samples. makes SPME an ideal sampling device. Additionally, headspace sam-
Water from simulated reghting was used in this study to calculate pling eliminates interference from the complex sample matrix and
the recovery of analytes and checking the impact of pyrolysis products has potential for eld analysis, making it even more attractive for
on analysis results and the SPME ber. In re investigation, HS-SPME use in routine testing.
has been mainly used for accelerants detection from solid material
(re debris). For the aqueous samples, headspace mode has not been Acknowledgments
reported so far. The HS-SPME method, compared to traditional LLE, limits
loss of hydrocarbons during the sampling procedure since the process in- This research work was supported by the system projects:
corporates extraction, concentration and injection into a single step, InnoDoktorant Scholarships for PhD students, IVth edition and
which, in turn, makes it more time efcient. Additionally, headspace sam- The development of interdisciplinary doctoral studies at the Gdansk
pling eliminates interference from the complex sample matrix, including University of Technology in modern technologies. Projects are
suspensions. Fig. 4 illustrates chromatograms obtained for an unltered co-nanced by the European Union in the frame of the European So-
water sample from reghting experiment. This illustrates that matrix cial Fund.

Fig. 4. Chromatograms obtained for water from reghting (a), spiked water from reghting at the concentration level 40 g L1 for each hydrocarbon (b) and spiked water from
reghting at the concentration level 200 g L1 for each hydrocarbon (c) using HS-SPMEGCFID; 1 = hexane, 2 = toluene, 3 = p-(m)-xylene, 4 = o-xylene, 5 = nonane, 6 =
decane, 7 = undecane, 8 = dodecane, 9 = tetradecane, 10 = pentadecane, 11 = hexadecane, 12 = eicosane.
M. Tankiewicz et al. / Microchemical Journal 108 (2013) 117123 123

Fig. 5. Chromatogram obtained for the contaminated tap water sample by 0.1% (v/v) of unleaded petrol using the HS-SPMEGCFID; 1 = hexane, 2 = toluene, 3 = p-(m)-xylene, 4 =
o-xylene, 5 = nonane, 6 = decane, 7 = undecane, 8 = dodecane.

The authors thank Mr. C. McGinness, University of the West of Scot- [20] R. Borusiewicz, J. Zieba-Palus, G. Zadora, The inuence of the type of accelerant,
type of burned materials, time of burning and accessibility of air on possibility
land, for technical assistance, and are grateful to Mrs. Ailsa Morrison for
of detection of the accelerant traces, Forensic Sci. Int. 160 (2006) 115126.
assistance with proof reading and constructive comments. [21] R. Borusiewicz, G. Zadora, J. Zieba-Palus, Application of head-space analysis with
passive adsorption for forensic purposes in the automated thermal desorption-gas
chromatographymass spectrometry system, Chromatographia 60 (2004) S133
S142.
References [22] M. Tankiewicz, J. Fenik, M. Biziuk, Solventless and solvent-minimized sample
preparation techniques for determining currently used pesticides in water
[1] E.S. Bodle, J.K. Hardy, Multivariate pattern recognition of petroleum-based samples: a review, Talanta 86 (2011) 822.
accelerants by SPME GCFID, Anal. Chim. Acta 589 (2007) 247254. [23] Ch. Fang, Y. Xiong, Q. Liang, Y. Li, P. Peng, Optimization of headspace single-drop
[2] J. DeHann, Kirk's Fire Investigation, fth ed. Prentice Hall, Upper Saddle River, microextraction technique for extraction of light hydrocarbons (C6C12) and its
New Jersey, 2002. potential applications, Org. Geochem. 42 (2011) 316322.
[3] W. Bertsch, Chemical analysis of re debris: was it arson? Anal. Chem. 68 (1996) [24] C.L. Arthur, J. Pawliszyn, Solid phase microextraction with thermal desorption
541A545A. using fused silica optical bers, Anal. Chem. 62 (1990) 21452148.
[4] United States Fire Administration Home Page, http://www.usfa.fema.gov(accessed [25] K. Furton, J. Almirall, J. Wang, J.C. Bruna, A novel method for the analysis of gaso-
28.06.2012). line from re debris using headspace solid-phase microextraction, J. Forensic Sci.
[5] Fire and Disaster Management Agency, Ministry of Internal Affairs and 41 (1996) 1222.
Communications, Japan, White Paper on Fire and Disaster Management, Gyousei, [26] K. Furton, J. Wang, Y.L. Walton, J. Almirall, The use of solid-phase microextraction/gas
Tokyo, 2005. chromatography in forensic analysis, J. Chromatogr. Sci. 38 (2000) 297306.
[6] . Stauffer, J.J. Lentini, ASTM standards for re debris analysis: a review, Forensic [27] J. Almirall, J. Wang, K. Lothridge, K. Furton, The detection and analysis of ignitable
Sci. Int. 132 (2003) 6367. liquid residues extracted from human skin using SPME/GC, J. Forensic Sci. 45
[7] American Society for Testing and Materials, ASTM E 1387-01 Standard test method for (2000) 453461.
ignitable liquid residues in extracts from re debris samples by gas chromatography, [28] K. Furton, J. Almirall, M. Bi, J. Wang, L. Wu, Application of solid-phase
ASTM, West Conshohocken, PA, 2002. microextraction to the recovery of explosives and ignitable liquid residues
[8] American Society for Testing and Materials, ASTM E 1618-01 Standard test from forensic specimens, J. Chromatogr. A 885 (2000) 419432.
method for ignitable liquid residues in extracts from re debris samples by gas [29] A. Steffen, J. Pawliszyn, Determination of liquid accelerants in arson suspected re de-
chromatographymass spectrometry, ASTM, West Conshohocken, PA, 2002. bris using headspace solid phase microextraction, Anal. Commun. 33 (1996) 129131.
[9] A.D. Pert, M.G. Baron, J.W. Birkett, Review of analytical techniques for arson residues, [30] J. Jackowski, The incidence of ignitable liquid residues in re debris as determined by
J. Forensic Sci. 51 (2006) 10331049. a sensitive and comprehensive analytical scheme, J. Forensic Sci. 42 (1997) 828832.
[10] B. Tan, J.K. Hardy, R.E. Snavely, Accelerant classication by GC/MS and multivariate [31] Q. Ren, W. Bertsch, A comprehensive sample preparation scheme for accelerants
pattern recognition, Anal. Chim. Acta 422 (2000) 3746. in suspect arson cases, J. Forensic Sci. 44 (1999) 504515.
[11] R.J. Mc Curdy, T. Atwell, M.D. Cole, The use of vapour phase ultra-violet spectroscopy [32] J.A. Lloyd, P.L. Edmiston, Preferential extraction of hydrocarbons from re debris
for the analysis of arson accelerants in re scene debris, Forensic Sci. Int. 123 (2001) samples by solid phase microextraction, J. Forensic Sci. 48 (2003) 130134.
191201. [33] American Society of Testing and Materials, ASTM Method E 2154-01 Standard prac-
[12] R.O. Keto, P.L. Wineman, Detection of petroleum-based accelerants in re debris tice for separation and concentration of ignitable liquid residues from re debris
by target compound gas chromatography/mass spectrometry, Anal. Chem. 63 samples by passive headspace concentration with solid-phase microextraction
(1991) 19641971. (SPME), ASTM, West Conshohocken, PA, 2002.
[13] K. Cavanagh, E.D. Pasquier, C. Lennard, Background interference from car carpets [34] A. Harris, J. Wheeler, GCMS of ignitable liquids using solvent-desorbed SPME for
the evidential value of petrol residues in cases of suspected vehicle arson, Forensic automated analysis, J. Forensic Sci. 48 (2003) 4146.
Sci. Int. 125 (2002) 2236. [35] K. Furton, J. Almirall, J. Wang, Variables controlling the recovery of ignitable liquid
[14] W. Bertsch, Volatiles from carpet: a source of frequent misinterpretation in arson residues from simulated re debris samples using solid phase microextraction/gas
analysis, J. Chromatogr. A 674 (1994) 329333. chromatography, Invest. Forensic Sci. Technol. 3576 (1999) 136141.
[15] American Society of Testing and Materials, ASTM Method E 1386-00 Standard [36] J. Alimirall, J. Bruna, K. Furton, The recovery of accelerants in aqueous samples
practice for separation and concentration of ignitable liquid residues from re de- from re debris using solid-phase microextraction (SPME), Sci. Justice 36
bris samples by solvent extraction, ASTM, West Conshohocken, PA, 2001. (1996) 283287.
[16] C. Huang, Y. Hsieh, Sample pretreatment of arson debris by supercritical uid ex- [37] J.Y. Horng, S.D. Huang, Determination of the semi-volatile compounds nitrobenzene,
traction, Huaxue 53 (1995) 139154. isophorone, 2,4-dinitrotoluene and 2,6-dinitrotoluene in water using solid-phase
[17] American Society of Testing and Materials, ASTM Method E 1385-00 Standard microextraction with a polydimethylsiloxane-coated ber, J. Chromatogr. A 678
practice for separation and concentration of ignitable liquid residues from re de- (1994) 313318.
bris samples by steam distillation, ASTM, West Conshohocken, PA, 2001. [38] C. Basheer, H.K. Lee, Hollow ber membrane-protected solid-phase microextraction
[18] J. Almirall, K. Furton, Characterization of background and pyrolysis products that of triazine herbicides in bovine milk and sewage sludge samples, J. Chromatogr. A
may interfere with the forensic analysis of re debris, J. Anal. Appl. Pyrolysis 71 1047 (2004) 189194.
(2004) 5167. [39] H. Yoshida, T. Kaneko, S. Suzuki, A solid-phase microextraction method for the
[19] American Society of Testing and Materials, ASTM Method E 1412-00 Standard detection of ignitable liquids in re debris, J. Forensic Sci. 53 (2008) 668676.
practice for separation and concentration of ignitable liquid residues from re de- [40] P. Konieczka, J. Namienik, Quality assurance and quality control in the analytical
bris samples by passive headspace concentration with activated charcoal, ASTM, chemical laboratorya practical approach, in: Ch.H. Lochmller (Ed.), Analytical
West Conshohocken, PA, 2001. Chemistry Series, CRC/Taylor & Francis, Boca Raton, 2009, pp. 131216.

You might also like