Journal of Chromatography A, 1217 (2010) 6816–6823

Contents lists available at ScienceDirect

Journal of Chromatography A
journal homepage: www.elsevier.com/locate/chroma

A cost effective, sensitive, and environmentally friendly sample preparation

method for determination of polycyclic aromatic hydrocarbons in solid samples
Chika Yamaguchi a , Wen-Yee Lee a,b,∗
a
Environmental Science and Engineering Ph.D Program, The University of Texas at El Paso (UTEP), 500 W. University Ave., El Paso, TX 79968-0513, USA
b
Department of Chemistry, UTEP, 500 W. University Ave., El Paso, TX 79968-0513, USA

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

Article history: A simple, cost effective, and yet sensitive sample preparation technique was investigated for deter-
Received 22 February 2010 mining Polycyclic Aromatic Hydrocarbons (PAHs) in solid samples. The method comprises ultrasonic
Received in revised form 10 August 2010 extraction, Stir Bar Sorptive Extraction (SBSE), and thermal desorption–gas chromatography–mass spec-
Accepted 23 August 2010
trometry to increase analytical capacity in laboratories. This method required no clean-up, satisfied PAHs
Available online 28 September 2010
recovery, and significantly advances cost performance over conventional extraction methods, such as
Soxhlet and Microwave Assisted Extraction (MAE). This study evaluated three operational parameters
Keywords:
for ultrasonic extraction: solvent composition, extraction time, and sample load. A standard material,
Ultrasonic extraction
Stir Bar Sorptive Extraction
SRM 1649 a (urban dust), was used as the solid sample matrix, and 12 priority PAHs on the US Envi-
SRM 1649a ronmental Protection Agency (US EPA) list were analyzed. Combination of non-polar and polar solvents
GC/MS ameliorated extraction efficiency. Acetone/hexane mixtures of 2:3 and 1:1 (v/v) gave the most satisfac-
Microwave Assisted Extraction tory results: recoveries ranged from 63.3% to 122%. Single composition solvents (methanol, hexane, and
Green Chemistry dichloromethane) showed fewer recoveries. Comparing 20 min with 60 min sonication, longer sonica-
tion diminished extraction efficiencies in general. Furthermore, sample load became a critical factor in
certain solvent systems, particularly MeOH. MAE was also compared to the ultrasonic extraction, and
results determined that the 20-min ultrasonic extraction using acetone/hexane (2:3, v/v) was as potent
as MAE. The SBSE method using 20 mL of 30% alcohol-fortified solution rendered a limit of detection
ranging from 1.7 to 32 ng L−1 and a limit of quantitation ranging from 5.8 to 110 ng L−1 for the 16 US EPA
PAHs.
© 2010 Elsevier B.V. All rights reserved.

1. Introduction is followed by condensation, clean-up, and enrichment prior to

instrumental analysis. However, the popular technique requires a
Polycyclic Aromatic Hydrocarbons (PAHs) comprise two or more high organic solvent volume, extensive extraction time (12–20 h),
fused aromatic rings, and physicochemical properties differ from and intensive manpower, resulting in a high PAH analysis cost and
one to another as chemical structure varies. As a result, each a conflict with the Green Chemistry concept as well. Alternative
PAH has discreet fate, transport, and distribution patterns in the techniques have been developed to overcome the impediments
environment [1]. Among over 100 different PAHs, the US Envi- in the traditional method [5,6]. For instance, Microwave Assisted
ronmental Protection Agency (US EPA) has identified 16 PAHs as Extraction (MAE) was applied to extract PAHs in airborne partic-
priority pollutants based on their health and environmental risk ulate matter, wood samples, marine sediment, muscle samples
concerns. Studying PAHs occurrence in environmental media has of polluted fish, and sewage sludge [7–11]. Accelerated Solvent
been focused on assessing human exposure to the pollutants and Extraction (ASE) was also utilized for extracting persistent organic
their impact on human health [2–4]; thus, measuring PAHs in the pollutants, such as PAHs, PCBs, and pesticides from soil, marine sed-
air, soil, and water attracted strong interests. iment, and urban dust [12–14]. The US EPA has approved the above
Soxhlet extraction is one of the major standard techniques two methods for extracting hydrophobic organic compounds from
for PAH determination in air or solid samples. The method solid materials (method 3545A and 2546). Indeed, the methods
uses dichloromethane or hexane/acetone mixture solvents and resolved the key issues of Soxhlet extraction, i.e. organic solvent
consumption and extraction time; however, the methods raised
another issue: a high capital cost. These environmentally friendly
∗ Corresponding author at: Department of Chemistry, UTEP, 500 W. University yet costly methods are, therefore, unavailable in many laboratories
Ave., El Paso, TX 79968-0513, USA. Tel.: +1 915 747 8413; fax: +1 915 747 5748. due to the capital cost. Consequently, science community promotes
E-mail address: wylee@utep.edu (W.-Y. Lee). alternative techniques that are inexpensive, readily available, and

0021-9673/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.chroma.2010.08.055
 C. Yamaguchi, W.-Y. Lee / J. Chromatogr. A 1217 (2010) 6816–6823 6817

easily used in many laboratories [15]. Ultrasonic extraction is one

of those techniques that fit the criteria and has gained a reputation
as a reliable method for extracting PAH from solid samples with
satisfactory recovery [16,17]. In this study, we further postulate
ultrasonic extraction using an ultrasonic bath for extracting PAHs
effectively from solid samples without clean-up procedures.
An ultrasonic bath is accessible in most laboratories, and ultra-
sonic extraction favorably deals with relevant factors to analytical
chemistry: capital and operating costs, environmental impact, and
level of automation [18]. Using an ultrasonic bath has an advan-
tage over a horn-type ultrasonic device described in EPA 3550C.
An ultrasonic bath allows a multiple sample extraction in one pro-
cess; thus, ultrasonic extraction using a bath significantly reduces
preparation time. Although ultrasonic extraction demonstrates less
extraction efficiency in some studies [16], the method was already
exploited for PAH extraction from solid samples [19–22]. Still,
relatively few studies have explored the extraction efficiency for
operational parameters such as solvent selection, sonication dura-
Fig. 1. Diagram of a conventional sample preparation process and the proposed
tion, and sample load. Therefore, it is essential to investigate these
method using SBSE.
parameters for the extraction efficiencies to determine the ultra-
sonic extraction potential and to provide meaningful comparison
between ultrasonic extraction and other extraction approaches. organic content, such as vegetable, baby food, coffee and peats,
This research also took Green Chemistry into consideration. SBSE was shown to have effective recovery of PAHs without or with
In the effort to reduce solvent and energy consumption and to minimal clean-up [30,15,17].
eliminate harmful organic solvents such as dichloromethane, a This study aimed to establish a simple, cost effective, environ-
solventless method called Stir Bar Sorptive Extraction (SBSE) was mentally friendly, and yet reliable technique for determining PAHs,
adopted in this study. SBSE is known for its high recovery and using an ultrasonic bath followed by SBSE–TDU–GC/MS with no
high extraction efficiency on hydrophobic and semivolatile organic clean-up process. Effects of solvent composition, duration of ultra-
pollutants including PAHs. SBSE has shown to be a rapid, environ- sonic extraction, and sample load were investigated to evaluate
mentally friendly, and reliable analytical technique [23–27]. The PAHs recoveries in solid samples.
fundamental SBSE methodology shares with solid phase microex-
traction (SPME). In the SBSE process, a Stir Bar (or TwisterTM ) coated 2. Experimental
with 50–300 ␮L of polydimethylsiloxane (PDMS) is placed in the
sample solution and is stirred for a pre-determined period. This 2.1. Materials
process allows the organic compounds to be extracted from the
solution phase to the Stir Bar. The equilibrium is controlled by Standard Material SRM 1649a (urban dust) was purchased
the partition coefficient between the PDMS phase and the solu- from the National Institute of Standards and Technology (NIST)
tion phase, KPDMS/w . The value of KPDMS/w increases with analyte’s (Gaithersburg, MD, USA). A mixture of PAH standards [phenan-
octanol–water partition coefficient (Kow ). Compounds with high threne (Phe), anthracene (Ant), fluoranthene (Ft), pyrene (Pyr),
Kow , such as PAHs, are prone to partition into the PDMS phase benz[a]anthracene (BaA), chrysene (Chry), benzo[b]fluoranthene
from the aqueous sample matrices; therefore, the PDMS phase on (BbFt), benzo[k]fluoranthene (BkFt), benzo[a]pyrene (BaP),
the Stir Bar enriches PAHs [23]. In general, a Stir Bar is placed indeno[1,2,3-cd]pyrene (InP), dibenz[a,h]anthracene (DahA), and
in a vial with a sample solution and stirred for a pre-determined benzo[ghi]perylene (BghiP) in methylene chloride/benzene (1:1)]
time ranging 0.5–24 h, depending on a sample matrix and a tar- was purchased from ULTRA Scientific (Kingstown, RI, USA). The
get compound. After stirring, the Stir Bar is removed from the mixture of internal standards, ASM-182 [1,4-dichlorobenzene-d4,
sample solution, and the adsorbed compounds are thermally des- acenaphthene-d10, chrysene-d12, naphthalene-d8, perylene-d12,
orbed from the PDMS phase in a thermal desorption unit (TDU) and and phenanthrene-d10 in dichloromethane/carbon disulfide (4:1)],
analyzed by gas chromatography and mass spectrometry (GC/MS). was purchased from AccuStandard (New Haven, CT, USA). Solvents
When a TDU is not available or thermally susceptible analytes are used in this study were of HPLC or higher grade: methanol: Burdick
targeted, liquid desorption with liquid chromatography or capil- & Jackson (Muskegon, MI, USA); acetone: VWR International (West
lary electrophoresis can be used as an alternative [28]. For example, Chester, PA); hexane: J. T Baker (Phillipsburg, NJ, USA); 2-propanol:
García-Falcón et al. reported to use SBSE to couple with HPLC (high Sigma-Aldrich (St. Louis, MO, USA); and dichloromethane: VWR
performance liquid chromatography) for the determination of free International.
PAHs in run-off water samples. After SBSE extraction, the Stir Bar
was placed in acetonitrile to allow PAHs desorption process to take 2.2. Extraction of PAHs from SRM1649a
place. Acetonitrile extracts were injected into the HPLC instrument
[29]. 2.2.1. Ultrasonic extraction
As illustrated in Fig. 1, conventional techniques often require Ultrasonic extraction was performed using an ultrasonic bath
a clean-up step prior to instrumental analysis, which may com- (Fisher Scientific FS 30H). To evaluate solvent effects, accurately
promise recovery of an analyte. In contrast, the SBSE method weighed SRM 1649a (100–150 mg) was placed into a 40 mL amber
potentially eliminates the clean-up step and minimizes solvent vial along with 40 mL of organic solvent. The solvents used
consumption; consequently, the method decreases the complexity were methanol (MeOH), hexane (HEX), dichloromethane (DCM),
and interferences in the sample preparation processes. In a review methanol/acetone (MeOH/ACE (1:1, v/v)), hexane/isopropanol
paper, David and Sandra detailed the SBSE principles and included (HEX/PrOH (1:1, v/v)), acetone/hexane (ACE/HEX (1:1, v/v) and
numerous SBSE applications for determination of PAH in aqueous (2:3, v/v)), dichloromethane/acetone/methanol (DCM/ACE/MeOH
samples without a clean-up step [25]. Even in samples with high (3:2:2, v/v/v)), and DCM followed by second extraction with
6818 C. Yamaguchi, W.-Y. Lee / J. Chromatogr. A 1217 (2010) 6816–6823

MeOH/ACE (1:1, v/v) (DCM–MeOH/ACE). The vial containing 40 mL 2.5. Determination of PAHs
of organic solvent with the sample was secured with a sili-
con cap and then sonicated in the ultrasonic bath for 20 or The retention time for individual PAHs and internal standards
60 min. After ultrasonic extraction, the extract was transferred to were determined by mass spectra using a scan mode prior to
a Kuderna–Danish (K–D) condenser. A solvent exchange was per- sample analysis. Identification of individual PAHs was based on
formed during the condensation to methanol or isopropanol when retention time comparison and mass-to-charge ratio (m/z). Ana-
non-alcohol base solvent was used for extraction. ASM-182 was lyte quantification was derived from analyte’s peak area. At least
added to the concentrated extracts after the condensation. The a five-point standard calibration was carried out within the range
extract’s final volume was adjusted to 10.0 mL by adding the same of 10–2000 ng L−1 for PAHs, and a 10–500 ng L−1 range was applied
alcohol used for the solvent exchange. The extracts were stored in to low level PAHs in the standard material. The r2 (r: regression
a refrigerator at 4 ◦ C until following SBSE procedure. Blank samples of coefficient) for PAHs varied between 0.9834 (DahA) and 0.999
were prepared for quality control purposes. (BaA and BeP). Although the US EPA 16 PAHs were analyzed, only
12 PAHs have certified values in SRM 1649a and were reported in
2.2.2. Microwave Assisted Extraction (MAE) this study.
MAE was performed with 100 mg of SRM 1649a and 40 mL of
a mixture of ACE/HEX (2:3, v/v). A CEM MARS Xpress Microwave
Accelerated Reaction System (CEM Corporation, Matthews, NC, 3. Results and discussion
USA) was used. The standard material and 40 mL of the solvent
were placed into a 100 mL Teflon® vessel. Microwave energy was 3.1. Effects of solvent composition
set at 600 W for the entire extraction process. Extraction was pro-
grammed based on the US EPA’s recommended conditions in the Solvent effect was studied for the following composi-
method 3546: rising to a final temperate of 110 ◦ C in 6 min and tion: methanol (MeOH), hexane (HEX), dichloromethane (DCM),
holding at 110 ◦ C for 14 min. The extracts were cooled, trans- methanol/acetone (MeOH/ACE (1:1, v/v)), hexane/isopropanol
ferred to K–D condensers, and concentrated to 10 mL; and solvent (HEX/PrOH (1:1, v/v)), acetone/hexane (ACE/HEX (1:1, v/v) and
exchange was carried out with the same manner described in Sec- (2:3, v/v)), dichloromethane/acetone/methanol (DCM/ACE/MeOH
tion 2.2.1. The extracts were stored at 4 ◦ C until subsequent SBSE (3:2:2, v/v/v)), and DCM followed by second extraction with
process. For quality control purposes, blank samples were pre- MeOH/ACE (1:1,  v/v) (DCM–MeOH/ACE). The recovered total 12
pared. PAHs [ PAHs = (Phe, Ant, Ft, Pyr, BaA, Chry, BbFt, BkFt, BaP,
InP, DahA, and BghiP)] in SRM 1649a ranged from 17.32 to
33.25 mg kg−1 with an average of 28.13 mg kg−1 . The values were
2.3. Stir Bar Sorptive Extraction (SBSE)
compared to the SRM certified value, 39.92 mg/kg, and individ-
ual and total PAHs recoveries were reported. The lowest and the
Extracts from ultrasonic extraction and MAE were enriched by
highest recovered PAHs were extracted in MeOH and ACE/HEX
the SBSE technique. The optimized SBSE condition that are 30%
(2:3) mixtures, respectively. Mixtures of polar and non-polar sol-
alcohol-fortified solution with 4-h stirring was adopted from our
vents had higher recoveries than single solvents in general: the
previous work [31]. In a 20 mL amber vial, an aliquot of 0.5 or 1.0 mL
results exhibited
 a similar trend reported by other studies [8,32,33].
of the extract from the ultrasonic extraction or the MAE was mixed
The order of PAHs recovered by various solvent compositions
with 14 mL of deionized (DI) water, and methanol or isopropanol
were MeOH < HEX < DCM < DCM/MeOH/ACE (2:3:3) < HEX/PrOH
was added to obtain a final volume of 20 mL at 30% alcohol content.
(1:1) < HEX/ACE (1:1) < MeOH/ACE (1:1) < HEX/ACE (3:2). PAHs
A commercially available Stir Bars (TwisterTM , 10 mm × 1 mm, Ger-
recoveries in various solvent systems were compared to the cer-
stel, Mülheim an der Ruhr, Germany) was placed in the vial, and the
tified values and shown in Fig. 2. Mixtures of polar and non-polar
solution was stirred for 4 h at 1000 rpm. The stir bar was removed
solvent showed higher recoveries than what was obtained using
from the solution, rinsed with DI water, dried with lint free paper,
single solvents, except DahA. The high recovery of DahA in hex-
and placed into a thermal desorption tube for GC–MS analysis.
ane could be due to the DahA’s low concentration in SRM 1649a.
DahA presented the lowest concentration (0.288 ± 0.02 mg kg−1 )
2.4. Thermal desorption–gas chromatography–mass in the standard material: it is significantly lower than InP
spectrometry (TD–GC/MS) (3.18 ± 0.72 mg kg−1 ) and BghiP (4.01 ± 0.91 mg kg−1 ). Therefore,
even with the same degree of variation in extract, the low
PAHs were analyzed by a thermal desorption unit, TDU (Ger- concentration of DahA will render a greater difference in the recov-
tel), coupled with a 6890 GC system and a 5973 N Mass Selective ery than those derived from InP or BghiP, which have higher
Detector (Agilent Technologies, Wilmington DE). The initial TDU concentrations. Further study on solvation parameter and other
temperature was 50 ◦ C. After holding for 0.5 min, the TDU temper- distribution properties would be needed to reach a conclusive
ature was increased to 300 ◦ C at 60 ◦ C min−1 and held for 5 min. discussion.
Desorption gas flow was set at 50 mL min−1 . During the desorption, Two sample Student’s t-tests (˛ = 0.05) were performed to com-
desorbed compounds were concentrated in a cold injection system, pare extraction efficiencies in sets of two solvent systems; Table 1
CIS-4 (Gerstel), at −40 ◦ C prior to GC injection. Once the desorption summarizes the result. A positive value denotes that the solvent
process was completed, the CIS temperature was ramped to 320 ◦ C on the left column exhibited higher extraction efficiency than that
at 12 ◦ C min−1 and held for 10 min in a solvent vent mode. on the top row for the number of PAHs. For example, to compare
Splitless mode was employed for the GC analysis. A ZB-5ms extraction efficiency between MeOH and ACE/HEX (2:3), one could
capillary column (30 m × 0.25 mm × 0.25 ␮m with 5% phenyl–95% look up MeOH on the left column and then find ACE/HEX (2:3) on
dimethylpolysiloxane, Phenomenex, USA) was used. The oven tem- the uppermost row. The value, −11, indicates that MeOH showed
perature was programmed as follows: held for 2 min at 50 ◦ C; raised significantly less efficiencies in extracting 11 PAHs against ACE/HEX
at 25 ◦ C min−1 to 150 ◦ C; increased at 3 ◦ C min−1 to 230 ◦ C; ramped (2:3). We postulate that the sum of each row expresses a provi-
at 8 ◦ C min−1 to 300 ◦ C; and held for 15 min at 300 ◦ C. The US EPA sional figure of the solvent system’s extracting efficiency. Based on
16 priority PAHs in samples were traced by Mass Selective detector the values, the least and the most effective solvent compositions
using selected ion mode (SIM). are MeOH and ACE/HEX. Closely examining the PAH recoveries in
 C. Yamaguchi, W.-Y. Lee / J. Chromatogr. A 1217 (2010) 6816–6823 6819

Fig. 2. Effect of solvent on extraction efficiency. Ultrasonic extraction for 20 min followed by 4-h SBSE at 1000 rpm. Error bars indicate the standard error (N = 3).

ACE/HEX (2:3) and ACE/HEX (1:1), both demonstrated no statisti- previously stated results: single solvents in ultrasonic extraction
cal difference for the 12 PAHs, yet PAHs were slightly different. A produced the least effective recovery for the PAHs, except for DahA.
Tukey-test was performed for the solvent systems (supplementary The solvent system combining three solvents, i.e. DCM/MeOH/ACE
data can be found in the Appendix). The Tukey-test warranted the (2:3:3), impaired extraction efficiency.

Table 1
The result of paired t-test (˛ = 0.05) for 12 individual PAHs (Phe, Ant, Ft, Pyr, BaA, Chry, BbFt, BkFt, BaP, InP, DahA, and BghiP) extracted by
different solvent systems. A positive value denotes that the solvent on the left column exhibited higher extraction efficiency than that on
 top row for the
the number of PAHs. A negative value indicates that the solvent on the left column is less effective for the number of PAHs.
PAHs (mg kg−1 ) is the total 12 PAHs extracted from NIST 1649a. In cells where two numbers are present, this indicates that the recovery
of individual PAH by the solvent system shows mixed outcome: significantly more efficient for certain PAHs but less so for others.
6820 C. Yamaguchi, W.-Y. Lee / J. Chromatogr. A 1217 (2010) 6816–6823

Table 2
Analysis of PAHs in NIST SRM 1649a (urban dust) (N = 3). Acetone/hexane (2:3, v/v) was used in ultrasonic extraction for 20 min; SBSE method was carried out in 30%
alcohol/water with 4 h extraction.

PAHs pKow Certified value Measured results % Recovery Literature results

(mg kg−1 ) Mean ± SD (mg kg−1 ) Mean ± SD Mean ± SD
(Karthikeyan et al.)

Phenanthrene 4.35 4.14 ± 0.37 3.43 ± 0.08* 82.3 ± 1.35 4.37 ± 0.34
Anthracene 4.35 0.432 ± 0.09 0.49 ± 0.08 109 ± 12.3 0.54 ± 0.01
Fluoranthene 4.93 6.45 ± 0.18 5.49 ± 0.40 85.5 ± 5.07 5.65 ± 0.41
Pyrene 4.93 5.29 ± 0.25 4.46 ± 0.24* 84.4 ± 3.74 4.95 ± 0.38
Benz[a]anthracene 5.52 2.208 ± 0.07 1.66 ± 0.15 71.2 ± 0.85 2.17 ± 0.16
Chrysene 5.52 3.049 ± 0.06 2.31 ± 0.04* 75.1 ± 0.57 3.44 ± 0.44
Benzo[b]fluoranthene 6.11 6.45 ± 0.64 7.85 ± 0.76 125 ± 7.76 6.43 ± 0.44
Benzo[k]fluoranthene 6.11 1.913 ± 0.17 1.85 ± 0.04 95.9 ± 1.18 1.50 ± 0.11
Benzo[a]pyrene 6.11 2.509 ± 0.09 2.24 ± 0.08* 88.3 ± 2.01 2.21 ± 0.17
Indeno[1,2,3-cd]pyrene 6.70 3.18 ± 0.72 2.00 ± 0.33* 60.5 ± 7.72 3.86 ± 0.76
Dibenz[a,h]anthracene 6.70 0.288 ± 0.02 0.28 ± 0.04 95.3 ± 9.83 0.34 ± 0.10
Benzo[ghi]perylene 6.70 4.01 ± 0.91 2.56 ± 0.47* 61.0 ± 8.76 3.42 ± 0.26

Karthikeyan et al. [8].

*
Statistically different from certified value.

The 20 min ultrasonic extraction efficiency using ACE/HEX (2:3) isopropanol was performed during K-D condensation for ACE/HEX
is shown in Table 2. The table also includes corresponding values (2:3).
reported by Karthikeyan et al. using a low temperature MAE and Two extraction periods, 20 and 60 min, were compared, and the
ACE/HEX (1:1) [8]. Statistical analysis suggests that the amounts results for MeOH/ACE (1:1) and ACE/HEX (2:3) are presented in
of Ant, Flt, BaA, BbFt, BkFt, and DahA extracted with our method Fig. 3 (MeOH is not shown). Statistical analyses found that no signif-
did not significantly differ from the corresponding certified val- icant difference exists between 20 min and 60 min extraction time
ues while other PAH recoveries are significantly fewer. Overall, for the three solvent compositions, except for InP in ACE/MeOH
ACE/HEX (2:3) and ACE/HEX (1:1) are suitable for PAH extraction. (1:1) and for Chry and BkFt in ACE/HEX (2:3). In these cases, 60 min
The PAH recoveries using ACE/HEX (2:3) solvent system ranged extraction time provided higher recoveries for the PAHs indicated.
from 60.5% to 125.7%. Taking into consideration of efficiency, 20 min sonication suffices
PAHs extraction from the solid material.
3.2. Effects of ultrasonic extraction time
3.3. Effects of extraction method
The extraction time effect was studied for three solvent sys-
tems: MeOH, MeOH/ACE (1:1), and ACE/HEX (2:3). ACE/HEX (2:3) MAE is a well-studied technique for PAHs extraction due to
previously exhibited the highest recovery; therefore, this solvent its low quantity of organic solvent consumption and its high
system was also selected in this section. MeOH/ACE (1:1) and efficiency [35]. MAE using ACE/HEX (2:3) was compared to
MeOH were also chosen to simplify the sample preparation pro- the ultrasonic extraction. The results of statistical analyses and
cess using the SBSE. During SBSE process, methanol is added to recoveries are shown in Table 3 and Fig. 4, respectively. Sta-
sample matrices to prevent adsorption of non-polar organic com- tistically significant difference was not observed between MAE
pounds onto the glass wall in many studies [25,34]. By using and sonication method. Slightly higher recovery was obtained
methanol or a mixture of methanol and acetone as an extraction by MAE for higher molecular weight PAHs, i.e. BkFt, BaP, InP,
solvent, solvent exchange process can be eliminated; therefore, DahA, and BghiP, although the sonication method demonstrated
using MeOH/ACE (1:1) or MeOH has an advantage to simplify better recoveries for lower molecular weight PAHs to some
the whole process. The sample was 100 mg of SRM 1649a, and extent. The results suggest that the more volatile PAHs are
40 mL of solvent was used in extraction. A solvent exchange to subject to poor recovery with MAE. This could be explained

Fig. 3. Effect of ultrasonic extraction periods, 20 min and 60 min, on extraction efficiency. Error bars indicate the standard error (N = 3).
 C. Yamaguchi, W.-Y. Lee / J. Chromatogr. A 1217 (2010) 6816–6823 6821

Table 3
Paired t-test (˛ = 0.05) results for 12 PAHs recovery. Ultrasonic extractions in various solvent systems and
MAE. MAE: results from the MAE performed in ACE/HEX (2:3, v/v); 20 min: results from the ultrasonic
extraction for 20 min in ACE/HEX (2:3, v/v); 60 min: results from the ultrasonic extraction for 60 min
in ACE/HEX (2:3, v/v); 2-step: ultrasonic extraction by−1DCM and followed by a second extraction using
MeOH/ACE (1:1, v/v) each for 20 min. PAHs (mg kg ) is the total 12 PAHs from NIST 1649a.

by their evaporation into the headspace during the heating were examined. As illustrated in Fig. 5, statistical analysis verified
process in MAE [35,36], resulting in the loss of these com- a significant difference (˛ = 0.05) between 20 and 100 mg of sam-
pounds. ple load. The smaller sample load, i.e. 20 mg, showed significantly
DCM extraction followed by a second ultrasonic extraction with higher recoveries for Ft, Pyr, BaA, BkFt, BaP, InP, DahA, and BghiP. On
MeOH/ACE (1:1) was studied to compare if the two-step extraction the other hand, the longer extraction time had a negative impact
(DCM–MeOH/ACE) has an advantage over single or mixture solvent for BaA, BaP, and DahA. Yet 100 mg of samples in ACE/HEX (2:3)
systems. A low recovery was observed for higher molecular weight demonstrated a higher extracting efficiency than 20 mg sample in
PAHs. Lower molecular weight PAHs, such as Phe and Pyr, were MeOH.
more effectively recovered. Based on the relative standard devi- The average of 12 PAHs’ recoveries for the 100 mg-ACE/HEX
ation, the greater data discrepancy was apparent in the two-step (2:3) was 86% while that for the 20 mg-MeOH system showed
process when compared to any other methods. 65%. Additionally, ACE/HEX (2:3) solvent system satisfactory
Wang et al. reported that MAE extraction was as effective as dealt with a higher sample load ranging 130–150 mg with
Soxhlet [37]; therefore, the results suggest that the ultrasonic no statistically significant difference (data not shown). In con-
extraction method is also comparable to Soxhlet when the same clusion, MeOH is deficient in extracting PAHs from SRM
solvent load is applied. Considering the capital cost, the ultrasonic 1649a.
extraction method provides compatible results for the recoveries
of PAHs to MAE extraction.
3.5. LOD and LOQ in SBSE

3.4. Effects of sample load As mentioned previously, SBSE coupled with thermal desorp-
tion has shown sufficient recovery and high extraction efficiency
Sample load effect on PAHs recovery was studied for 20 and on PAHs. The limits of detection (LOD) and the limits of quantifica-
100 mg sample amounts using 40 mL of MeOH. The sample to sol- tion (LOQ) for PAHs in SBSE–TD–GC–MS were tested. Using a 30%
vent ratio selected was comparable to the value (1 g of sample in alcohol-fortified solution and a final volume of 20 mL, spiked PAH
200 mL of solvent) reported in the NIST Certificate of Analysis of solutions were extracted by a Stir Bar with 4-h stirring time at
SRM 1649a. Although methanol was incompetent among previ- 1000 rpm. Because the solution in the SBSE procedure was fixed
ously studied solvent systems, using methanol offers a significant volume (20 mL), LODs are reported in ng L−1 . Analyzed by GC/MS,
advantages of simple extraction process; therefore, methanol was LOD and LOQ values were determined by a signal-to-noise ratio
studied anew for this section. Extraction times of 20 and 60 min of three-to-one and ten-to-one, respectively. As listed in Table 4,

Fig. 4. Effect of extraction method on extraction efficiency. MAE ACE/HEX (2:3): MAE extraction with ACE/HEX (2:3); Sonic 20 ACE/HEX (2:3): sonication for 20 min with
ACE/HEX (2:3); Sonic 60 ACE/HEX (2:3): sonication for 60 min with ACE/HEX (2:3); DCM–MeOH/ACE (1:1): sonication for 20 min with DCM followed by additional 20 min
with MeOH/ACE (1:1). All methods were followed by 4-h SBSE at 1000 rpm. Error bars indicate the standard error (N = 3).
6822 C. Yamaguchi, W.-Y. Lee / J. Chromatogr. A 1217 (2010) 6816–6823

Fig. 5. Effect of sample load on extraction efficiency using methanol as extraction solvent. Error bars indicate the standard error (N = 3).

Table 4 ration techniques, such as Soxhlet and MAE, ultrasonic extraction

Limit of detection (LOD) and limit of quantification (LOQ) in ng L−1 for investigated
is fast and effective. The postulated method requires only a
PAHs in 20 mL of SBSE solution followed by GC/MS. Method Detection Limit (MDL)
was calculated based on 20 mg of sample load. 20 min extraction and a common device in laboratories; this
method significantly improves cost and extraction time over 2–4 h
PAHs LOD (ng L−1 ) LOQ (ng L−1 ) MDL (␮g kg−1 )
for MAE (including the cooling time) and 12–24 h for Soxhlet
Naphthalene 31.9 106 extraction.
Acenaphthylene 3.59 12.0 The solvent effect study demonstrated that combination of non-
Acenaphthene 3.07 10.2
polar and polar solvent ameliorate extracting PAHs from SRM
Fluorene 2.65 8.83
Phenanthrene 7.01 23.4 8.52 1649a. Acetone/hexane mixtures at 2:3 and 1:1 (v/v) ratios gave
Anthracene 1.73 5.77 1.59 the most satisfactory results for PAHs recovery from SRM 1649a.
Fluoranthene 4.79 16.0 5.60 Longer sonication duration diminished extraction efficiencies in
Pyrene 4.33 14.4 5.13
most cases; therefore, 20 min extraction suffices PAHs extraction.
Benz[a]anthracene 1.98 6.60 2.78
Chrysene 4.65 15.5 6.19 The comparison of the two extraction methods, ultrasonic extrac-
Benzo[b]fluoranthene 20.9 69.7 16.7 tion and MAE, rendered the similar result. The ultrasonic sonic
Benzo[k]fluoranthene 9.40 31.3 9.80 extraction gave higher recoveries on PAHs that are more volatile
Benzo[a]pyrene 5.23 17.4 5.92 while MAE was more effective for less volatile PAHs. In addition,
Indeno[1,2,3-cd]pyrene 5.15 17.2 8.51
sample load becomes a critical factor on extraction efficiency when
Dibenz[a,h]anthracene 4.66 15.5 4.89
Benzo[ghi]perylene 7.09 23.6 11.6 methanol is used. Regardless, the solvent system is still more cru-
cial than sample load. LOD ranged from 1.7 ng L−1 for anthracene
to 32 ng L−1 for naphthalene, and LOQ ranged from 5.8 ng L−1 for
the LOD for PAHs ranged between 1.7 and 32 ng L−1 , and the LOQ anthracene to 110 ng L−1 for naphthalene under the SBSE condi-
ranged from 5.8 to 110 ng L−1 for the US EPA 16 PAHs. García-Falcón tions: 20 mL of 30% alcohol-fortified solution with 4 h stirring at
et al. reported similar analytical capacity of using SBSE to determine 1000 rpm. Although SBSE was utilized in this study, SBSE is not a
PAHs in water [29]. compulsory process for determining of PAHs: an orthodox clean-
Describing MDLs in ␮g kg−1 without citing sample amounts up process substitutes SBSE. Similarly, when SBSE is available but
used is frequently invalid for practical procedure [38]. This method not a TDU, solvent desorption may replace the TDU. The pre-
was optimized to use ACE/HEX (2:3) for 20–150 mg of solid sam- sented method is simple, fast, use lower volumes of inexpensive
ples. LODs in Table 4 stated the instrument’s sensitivity based on and non-halogenated solvents. It incorporated Green Chemistry in
alcohol-fortified samples in SBSE–TD–GC–MS for individual PAHs. environmental analysis, and presents to be particularly useful when
Recoveries express ultrasonic extraction efficiencies for these com- estimated level of PAHs is desired for screening purposes or for
pounds in SRM 1649a (Table 2). Method Detection Limit (MDL) preliminary study.
was readily estimated by using recovery and LOD [39] and was
summarized in Table 4. MDLs were determined by (LOD × final vol- Acknowledgements
ume)/(sample weight × extraction recovery). Upon specifying the
sample load at 20 mg and using 20 mL as the final volume, the MDLs The project was partially supported by Grant Number
in this particular study were estimated to be from 1.59 ␮g kg−1 S11ES013339 from the National Institute of Environmental
(Ant) to 16.7 ␮g kg−1 (BbFt). Health Sciences, EPA Student Support Assistance Agreement (T-
83056401), and NIH SCORE Program at the University of Texas at
4. Conclusions El Paso (UTEP) (Project Number 2S06GM008012-38). The content
is solely the responsibility of the authors and does not necessarily
This study demonstrated that ultrasonic extraction using an represent the official views of the funding agencies.
ultrasonic bath is a rapid, reliable, and environmental friendly The assistance provided by the department of Chemistry and the
approach in recovering PAHs from a solid matrix. Ultrasonic extrac- Center for Environmental Resource Management at UTEP is grate-
tion using ACE/HEX (2:3 and 1:1, v/v) solvents provided sufficient fully recognized. Drs. Bonnie M. Gunn and Elaine F. Fredericksen
extraction efficiencies. Compared to commonly used sample prepa- are thanked for providing valuable feedback on the manuscript. Dr.
 C. Yamaguchi, W.-Y. Lee / J. Chromatogr. A 1217 (2010) 6816–6823 6823

Julia Bader in Statistical Consulting Laboratory at UTEP is acknowl- [16] L. Rey-Salgueiroa, X. Pontevedra-Pombalb, M. Álvarez-Casasa, E. Martínez-
edged for her help in statistical analysis. Carballoa, M.S. García-Falcóna, J. Simal-Gándara, J. Chromatogr. A 1216 (2009)
5235.
[17] L. Rey-Salgueiro, E. Martínez-Carballo, M.S. García-Falcón, C. González-
Appendix A. Supplementary data Barreiro, J. Simal-Gándara, Food Chem. 115 (2009) 814.
[18] J.R. Dean, G. Xiong, TrAC, Trends Anal. Chem. 19 (2000) 553.
[19] D.R. Banjoo, P.K. Nelson, J. Chromatogr. A 1066 (2005) 9.
Supplementary data associated with this article can be found, in [20] J.P. Bossio, J. Harry, C.A. Kinney, Chemosphere 70 (2008) 858.
the online version, at doi:10.1016/j.chroma.2010.08.055. [21] E. Martinez, M. Gros, S. Lacorte, D. Barceló, J. Chromatogr. A 1047 (2004)
181.
[22] A. Tor, M.E. Aydin, S. Özcan, Anal. Chim. Acta 559 (2006) 173.
References [23] E. Baltussen, P. Sandra, F. David, C. Cramers, J. Microcolumn Sep. 11 (1999)
737.
[1] M.M.C. Ferreira, Chemosphere 44 (2001) 125. [24] E. Baltussen, C. Cramers, P. Sandra, Anal. Bioanal. Chem. 373 (2002) 3.
[2] Y. Hu, Z. Bai, L. Zhang, X. Wang, L. Zhang, Q. Yu, T. Zhu, Sci. Total Environ. 382 [25] F. David, P. Sandra, J. Chromatogr. A 1152 (2007) 54.
(2007) 240. [26] E. Pérez-Carrera, V.M.L. León, A.G. Parra, E. González-Mazo, J. Chromatogr. A
[3] C.-E. Boström, P. Gerde, A. Hanberg, B. Jernström, C. Johansson, T. Kyrklund, 1170 (2007) 82.
A. Rannug, M. Törnqvist, K. Victorin, R. Westerholm, Environ. Health Perspect. [27] O. Alvarez-Avilés, L. Cuadra-Rodríguez, F. González-Illán, J. Quiñones-González,
110 (2002) 451. O. Rosario, Anal. Chim. Acta 597 (2007) 273.
[4] R.J. Delfino, Environ. Health Perspect. 110 (Suppl. 4) (2002) 573. [28] A. Prieto, O. Basauri, R. Rodil, A. Usobiaga, L.A. Fernández, N. Etxebarria, O.
[5] N.a. Saim, J.R. Dean, M.P. Abdullah, Z. Zakaria, J. Chromatogr. A 791 (1997) Zuloaga, J. Chromatogr. A 1217 (2010) 2642.
361. [29] M.S. García-Falcón, C. Pérez-Lamela, J. Simal-Gándara, Anal. Chim. Acta 508
[6] L. Ramos, J.J. Ramos, U.A.T. Brinkman, Anal. Bioanal. Chem. 381 (2005) 119. (2004) 177.
[7] V. Pino, J.H. Ayala, A.M. Afonso, V. González, J. Chromatogr. A 869 (2000) 515. [30] P. Sandra, B. Tienpont, F. David, J. Chromatogr. A 1000 (2003) 299.
[8] S. Karthikeyan, R. Balasubramanian, S.W. See, Talanta 69 (2006) 79. [31] R.J. De La Torre-Roche, W.-Y. Lee, S.I. Campos-Díaz, J. Hazard. Mater. 163 (2009)
[9] L. Pensado, C. Casais, C. Mejuto, R. Cela, J. Chromatogr. A 869 (2000) 505. 946.
[10] T. Pena, L. Pensado, C. Casais, C. Mejuto, R. Phan-Tan-Luu, R. Cela, J. Chromatogr. [32] C. Domeño, M. Blasco, C. Sánchez, C. Nerín, Anal. Chim. Acta 569 (2006)
A 1121 (2006) 163. 103.
[11] P. Villar, M. Callejón, E. Alonso, J.C. Jiménez, A. Guiraúm, Anal. Chim. Acta 524 [33] M.T.O. Jonker, A.A. Koelmans, Environ. Sci. Technol. 36 (2002) 4107.
(2004) 295. [34] B. Kolahgar, A. Hoffmann, A.C. Heiden, J. Chromatogr. A 963 (2002) 225.
[12] P. Popp, P. Keil, M. Möder, A. Paschke, U. Thuss, J. Chromatogr. A 774 (1997) [35] V. Camel, TrAC, Trends Anal. Chem. 19 (2000) 229.
203. [36] M. Letellier, H. Budzinski, L. Charrier, S. Capes, A.M. Dorthe, Fresenius’ J. Anal.
[13] K. Li, M. Landriault, M. Fingas, M. Llompart, J. Hazard. Mater. 102 (2003) Chem. 364 (1999) 228.
93. [37] W. Wang, B. Meng, X. Lu, Y. Liu, S. Tao, Anal. Chim. Acta 602 (2007) 211.
[14] S. Tao, Y.H. Cui, F.L. Xu, B.G. Li, J. Cao, W.X. Liu, G. Schmitt, X.J. Wang, W.R. Shen, [38] A.B. Fialkov, U. Steiner, S.J. Lehotay, A. Amirav, Int. J. Mass Spectrom. 260 (2007)
B.P. Qing, R. Sun, Sci. Total Environ. 320 (2004) 11. 31.
[15] M.S. Garcia-Falcon, B. Cancho-Grande, J. Simal-Gandara, Food Chem. 90 (2005) [39] R.C. Prados-Rosales, J.L. Luque García, M.D. Luque de Castro, J. Chromatogr. A
643. 993 (2003) 121.