Journal of Chromatography A, 1167 (2007) 171–177

Analysis of polynuclear aromatic hydrocarbons in heavy

products derived from coal and petroleum by high
performance liquid chromatography
Changming Zhang ∗ , Xiaohang Zhang,
Jianli Yang, Zhenyu Liu
State Key Laboratory of Coal Conversion, Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, China
Received 13 October 2006; received in revised form 26 May 2007; accepted 23 August 2007
Available online 26 August 2007

Abstract
A systematic study was made on the identification and quantitative determination of polynuclear aromatic hydrocarbons (PAHs) in heavy products
derived from coal and petroleum with high performance liquid chromatography (HPLC). After the separation of PAHs by a high-resolution column,
identification was made through four methods: (1) the relative retention time (RRT) method, (2) the stop-flow-UV scanning method, (3) the method
of fluorescence characteristic index Φ and (4) the method of V index at different UV wavelengths. For the quantitative determination of the
components, methods of external standard (E-X), internal standard (I-N) and external standard-response factors (E-F) were compared. The E-F
method was recommended by the present paper. For the determination of quantitative response factors (F) two methods were studied, including
the HPLC peak-area method (Peak-area method) and the UV absorbance method using a UV spectrometer (absorbance method). The absorbance
method was better and is recommended by the present paper. The F values of 30 PAHs from the two different methods are given. The samples
analyzed include a coal tar pitch, a thermal cracking residue oil and a residual oil from ethylene production.
© 2007 Elsevier B.V. All rights reserved.

Keywords: High performance liquid chromatography; Polynuclear aromatic hydrocarbons; Identification and quantitative determination

1. Introduction anthracene (99.14% pure), phenanthrene (96.28% pure) and

As notorious pollutants and valuable raw materials for
the manufacture of chemicals (such as decahydronaphthalene,
phenanthrenequinone, anthraquinone, anthralin, etc.), PAHs
have long been important subjects of analysis studies. Numer-
ous reports on the determination of PAHs in the environment
by HPLC have been published [1–7]. But new reports on the
analysis of PAHs by HPLC in the heavy products of coal and
petroleum industries, which are the main sources of PAHs are
relatively few [8–13].
Besides being a powerful tool for PAH analysis, HPLC is
also indispensable for the development of coal and petroleum
derivative utilization technologies. For example, with the moni-
toring of HPLC analysis, established by the authors, high purity
∗ Corresponding author. Fax: +86 351 404 8571.
E-mail address: zhangcm@sxicc.ac.cn (C. Zhang).
carbazole (98.88% pure) [14] have been produced commer-
cially in Jinfeng Chemical Factory, Shanxi, China, at a scale
of 1000 tonnes per year.
However, to meet the needs of further development in deep
processing technologies of coal and petroleum derivatives,
higher requirements are demanded in HPLC analysis. But lit-
eratures on PAHs in heavy oil are too few to meet the demand.
For qualitative analysis, only a few PAH components can be
identified. For quantitative analysis, the situation is even worse.
Facing the complexity and large number of PAH components
in samples, even preliminary results of qualitative and quan-
titative analysis are often unreachable. The objective of this
work is to establish a systematic and integrated method of PAH
analysis by HPLC to meet the need of researches in heavy oil
utilization.
For this purpose, the present work carried out a systematic
study and comparison of different methods and provided practi-
cal data for reference. We hope that our work may be helpful for

0021-9673/$ – see front matter © 2007 Elsevier B.V. All rights reserved.
doi:10.1016/j.chroma.2007.08.048
172 C. Zhang et al. / J. Chromatogr. A 1167 (2007) 171–177

Table 1
Name and source of practical samples
No. Name Source

1 Thermal cracking residual oil Daqing Refining Factory

2 Residual oil from Dong Fang Hong Refining Factory.
ethylene production
3 Coal tar pitch Taiyuan Iron and Steel
Industry Company

the analysis of PAHs and the development of new hydrocarbon

processing technologies.

2. Experimental

2.1. Apparatus, reagents and samples

The HPLC instrument used was a Shimadzu LC-3A high

performance liquid chromatograph, equipped with a SPD-1 Fig. 2. The chromatogram of a petroleum (Daqing) thermal cracking residual oil.
multi-wavelength UV detector (Shimadzu), a UVD-2 UV detec- Chromatographic conditions: Shimadzu LC-3A HPLC chromatograph. Detec-
tor (λ254 nm), and an RF-510 fluorescence detector (Shimadzu). tors: RF-510 fluorescence detector and UVD-2 ultraviolet detector (in series).
Column: 0.46 cm × 25 cm, packed with Zorbax-ODS, 5 ␮m. Column tempera-
The instrument used to determine UV absorbance was a Shi- ture: 19 ◦ C. Mobile phase: methanol/water = 85/15 (v/v), Flow rate: 1.0 ml/min.
madzu UV-310 ultraviolet spectrometer.
Methanol as the mobile phase was an analytical grade reagent Identification of peaks:
produced by Tianjin Chemical Reagent Factory (China). Pure
(1) Methylnaph- (2) 2,3-Dimethyl- (3) Acenaphthene
thalene naphthalene
(4) Phenanthrene (5) Anthracene (6) 2,6-Dimethylnaphthalene
(7) p-Terphenyl (8) Fluoranthene (9) 1-Methylanthracene
(10) Pyrene (11) m-Terphenyl (12) 2.3-Benzfluorene
(13) Benz[a]anth- (14) Chrysene (15) Benz[j]fluoranthene
racene
(16) Benzo[e]pyr- (17) Perylene (18) Benz[k]fluoranthene
ene
(19) Benzo[a]pyr- (20) 7,12-Dimeth-
ene ylbenz[a]anthracene
(21) 1,3,5-Triphe- (22) Coronene
nylbenzene

PAH reagents were supplied by Aldrich Chemical Company

(USA), including anthracene (98% pure), benzo[a]pyrene (98%,
pure), phenanthrene (98% pure), Chrysene (99% pure), fluoran-
thene (99% pure), pyrene (98% pure), etc.
The names and sources of the samples are listed in Table 1.
Fig. 1. Separation of a mixture of pure PAHs. Chromatographic conditions:
Shimadzu LC-3A HPLC chromatograph. Detectors: RF-510 fluorescence detec- 2.2. HPLC analysis
tor and UVD-2 ultraviolet detector (in series). Column: 0.46 cm × 25 cm
packed with Zorbax-ODS, 5 ␮m. Column temperature: 19 ◦ C. Mobile phase:
2.2.1. Optimum chromatographic conditions
methanol/water = 85/15 (v/v). Flow rate: 1.0 ml/min.
After careful studies, the optimum chromatographic condi-
Peaks:
tions are determined as the following: column packing: Zorbax-
ODS, particle size 5 ␮m; column size: 250 mm × 4.6 mm (i.d.);
(1) Benzene (2) Toluene (3) Naphthalene mobile phase: methanol/water = 85/15–90/10 (v/v, %); con-
(4) Acenaphthylene (5) Biphenyl (6) ␤-Methylnaphthalene
(7) Acenaphthene (8) Phenanthrene (9) Anthracene
centration of sample solution: 200–510 ␮g/ml; injection size:
(10) 2,6-Dimethyl- (11) Fluoranthene 0.2–4 ␮l; column temperature: 19–22 ◦ C.
naphthalene
(12) Pyrene (13) m-Terphenyl (14) Chrysene 2.2.2. Qualitative analysis
(15) Perylene (16) Benzo[a]pyrene
Four methods were used for qualitative analyses:
(17) 7,12-Dimethyl- (18) 1,3,5-Triphenylbenzene
benz[a]anthracene
(1) The method of relative retention time (RRT)
 C. Zhang et al. / J. Chromatogr. A 1167 (2007) 171–177 173

component x, either pure reagent or in samples, in the same

run.
By comparing the Φ s of the standards with those of the
samples, the components can be identified.
(4) The method of UV characteristic index V
In this method, two UV detectors of different wavelengths
were connected in series, with one of them set at 254 nm.
The index V is defined as

Sxλ
V = (2)
Sx254

where Sxλ is the UV response at UV wavelengths λ of com-

ponent x, either pure reagent or in samples, and Sx254 is the
UV response at 254 nm of component x, either pure reagent
or in samples in the same run.

Fig. 3. The chromatogram of a residual oil from ethylene production. Chro-

matographic conditions: Shimadzu LC-3A HPLC chromatograph. Detectors:
RF-510 fluorescence detector and UVD-2 ultraviolet detector (in series). Col-
umn: 0.46 cm × 25 cm, packed with Zorbax-ODS, 5 ␮m. Column temperature:
20 ◦ C. Mobile phase: methanol/water = 90/10 (v/v). Flow rate: 0.9 ml/min.
By comparing the V of the standards with those of the compo-
nents in samples, the identities of the composition can be further
verified.

Identification of peaks:

(1) Naphthalene (2) Biphenyl (3) ␤-Methylnaphthalene

(4) Acenaphthene (5) Phenanthrene (6) Anthracene
(7) 2,6-Dimethylna- (8) Fluoranthene (9) 9-Methylanthracene
phthalene
(10) Pyrene (11) 2-Methylanthracene (12) Benz[a]anthracene
(13) Chrysene (14) 9,10-Dimethylathracene (15) Benz[j]fluoranthene
(16) Benzo[e]py- (17) Perylene (18) Benzo[a]pyrene
rene

This method is one of the methods widely used in the

identification of components in chromatographic analysis.
In this method, the identities of components were checked
by RRT values of pure PAHs. In this paper, phenanthrene
was used as the internal standard.
(2) The method of stop-flow UV scanning
This method is based on the fact that different PAHs have
different UV spectra. Though it is effective, the practice is Fig. 4. The chromatogram of a petroleum ether extract of coal tar pitch. Chro-
matographic conditions: Shimadzu LC-3A HPLC chromatograph. Detectors:
laborious. During the analysis, the flow of the mobile phase UVD-2 ultraviolet detector, ASD-1 ultraviolet detector (in series). Column:
was stopped at the peak height of an effluent component, 0.46 cm × 25 cm, packed with Zorbax-ODS, 5 ␮m. Column temperature: 20 ◦ C.
and UV scanning was carried out. The UV spectrum of the Mobile phase: methanol/water =/15 (v/v). Flow rate: 1.3 ml/min.
component was then compared with the standard UV spectra
of different pure PAHs for identification. Identification of peaks:
(3) The method of fluorescence characteristic index Ф (1) Naphthalene (2) Acenaphthylene
In this method, the UV and fluorescence detectors were (3) ␤-Methylnaphthalene (4) Fluorene (5) Phenanthrene
connected in series and the fluorescence responses at spe- (6) Anthracene (7) p-Terphenyl (8) Fluoranthene
cific excitation and emission wavelengths were compared (9) 1-Methylathracene (10) Pyrene (11) m-Terphenyl
(12) 2,3-Benzfluorene (13) Benz[a]anthracene (14) Chrysene
with reference to a specific UV response. The index Φ is (15) 9,10-Dimet- (16) Benz[j]fluoranthene (17) Benzo[e]pyrene
defined as hylanthracene
(18) Perylene (19) Benz[k]fluoranthene (20) Benzo[a]pyrene
Sxf
Φ= (1) (21) 7,12-Dimethylbe-
Sxu nz[a]anthracene
(22) 1,3,5-Triphenylben- (23) Coronene
where Sxf is the fluorescence response of component x, either zene
pure reagent or in samples, and Sxu is the UV response of
174 C. Zhang et al. / J. Chromatogr. A 1167 (2007) 171–177

Fig. 5. Comparison of UV spectra of fluoranthene and pyrene with the UV spectra of peaks (components) no. 8 and no. 10 in Fig. 4. A1-sample: peak no 8 of Fig. 4.
A2-pure reagent: fluoranthene. B1-sample: peak no 10 of Fig. 4. B2-pure reagent: pyrene.

2.2.3. Quantitative determination by the method of external were calculated by

standard-response factor (E-F)  
Sx /Sph
Products derived from coal and petroleum contains a great Cx % = Cph % × (3)
number of different PAHs, and their quantitative analysis with Fph /Fx
internal standards is practically impossible. But to determine where Cx % and Cph % are the contents of component x and
every component with their corresponding external standards phenanthrene, respectively, Sx and Sph are the peak areas of
(the E-X method) is very laborious. The author proposed to use component x and phenanthrene, respectively, and Fx and Fph are
a method that combines E-X with response factors (the E-F). In the response factors of component x and phenanthrene, respec-
the analysis, one of the components is first determined by E-X tively. In the present paper, the response factor of phenanthrene
method. Then, other components are determined quantitatively was taken as 1.
by their response factors relative to this component. In this work,
phenanthrene was chosen as the standard, and other components
2.2.4. Determination of response factor (F)
In using the E-F method, the determination of response factor
F is the key step. F can be determined by two methods.
Table 2
Qualitative analysis of residual oil from ethylene production by the Φ method (1) Chromatographic peak area method
PAH Φ Pure Φ Sample Deviation Ex (nm) Em (nm)
A PAHs mixture of known concentrations was prepared
and subjected to HPLC analysis. All of these measurements
Acenaphthene 0.11 0.11 0.00 250 380
Acenaphthene 1.01 1.11 0.10 265 370
Phenanthrene 0.14 0.12 0.02 250 380 Table 3
Phenanthrene 0.25 0.27 0.02 265 370 Qualitative analysis of “Daqing” thermal cracking residual oil by the Φ method
Anthracene 0.74 0.71 0.03 250 380 PAH Φ Pure Φ Sample Deviation Ex (nm) Em (nm)
Anthracene 0.01 0.05 0.04 265 370
Biphenyl 0.17 0.16 0.01 250 315 Acenaphthene 1.01 1.11 0.10 265 370
␤-Methylnaphthalene 0.06 0.06 0.00 250 315 Phenanthrene 0.25 0.27 0.02 265 370
Benzo[a]pyrene 3.40 3.00 0.40 380 450 Anthracene 0.01 0.05 0.04 265 370
Benzo[a]pyrene 3.50 3.10 0.40 380 460 Benzo[a]pyrene 3.50 3.10 0.40 380 460
Chrysene 1.45 1.42 0.03 265 370 Chrysene 1.45 1.42 0.03 265 370
Perylene 3.00 3.42 0.42 380 460 Perylene 3.00 3.22 0.22 380 460
Coronene 3.00 3.30 0.30 380 460 Coronene 3.00 3.30 0.30 380 460

UV λ = 254 nm. UV λ = 254 nm.

 C. Zhang et al. / J. Chromatogr. A 1167 (2007) 171–177 175

Table 4
V values of components in an petroleum ether extract of coal tar pitch compared with those of pure reagents
Peak no in Fig. 4 Compound Pure 290/254 (nm) Sample 290/254 (nm) Pure 270/254 (nm) Sample 270/254 (nm)

1 Naphthalene 0.909 0.951 1.042 0.734

2 Acenaphthylene 1.120 1.003 1.022 1.161
3 ␤-Methylnaphthalene 0.982 0.978 1.083 1.111
4 Fluorene 1.103 1.108 0.986 0.894
5 Phenanthrene 0.177 0.182 0.271 0.223
6 Anthracene 0.510 0.523 0.041 0.049
7 p-Terphenyl 1.010 0.988 1.008 1.428
8 Fluoranthene 0.998 0.810 0.981 0.774
9 1-Methylanthracene 1.580 1.770 0.921 0.437
10 Pyrene 0.502 0.513 2.279 1.796
11 m-Terphenyl 0.137 0.156 0.189 0.132
12 2.3-Benzfluorene 0.840 0.821 0.889 0.876
13 Benz[a]anthracene 1.010 1.092 1.050 1.043
14 Chrysene 0.232 0.108 1.145 1.153
15 9,10-Dimethylanthracene 2.133 2.139 1.234 1.306
16 Benz[j]fluoranthene 0.971 0.968 0.869 0.795
17 Benzo[e]pyrene 1.007 1.003 1.007 1.003
18 Perylene 0.826 0.820 0.788 0.792
19 Benz[k]fluoranthene 0.989 0.460 0.935 0.733
20 Benzo[a]pyrene 0.959 0.690 0.970 0.973
21 7,12-Dimethylbenz[a]anthracene 2.133 2.136 1.233 1.229
22 1,3,5-Triphenyl-benzene 0.072 0.069 0.554 0.539
23 Coronene 1.135 1.130 1.105 1.110

should be performed at the same conditions. The F factors
of each component were calculated from their peak areas
by the following equation:
[Sx /Wx %]
Fx =
[Sph /Wph %]
where Fx is the F value of component x, Sx and Sph are peak
areas of component x and phenanthrene, respectively, and
Wx % and Wph % are the percentage contents of component
x and phenanthrene.
(2) UV absorbance method
PAH solutions of known concentrations were prepared
and the UV absorbance of each hydrocarbon was determined
by a UV spectrometer. At the same wavelength (λ254 nm),
F values were calculated from the following equation
[Ax /Wx %]
Fx =
[Aph /Wph %]
where Fx is the F value of component x, Ax and Aph are the
UV absorbance of component x and phenanthrene, respec-
tively, and Wx % and Wph % are the known concentrations of
component x and phenanthrene, respectively.
3. Results and discussions
3.1. Separation of PAHs
By applying a high-resolution chromatographic column
under a series of optimal chromatographic conditions, the reso-
lution of the critical pair of phenanthrene/acenaphthene was 5,
which is better than literature values [8–13].
Figs. 1–4 indicate high resolution for either pure PAHs or
practical samples. Compared with the analysis of similar sam-
ples [8–13], Fig. 3 is superior in clarity and resolution. The
same trend was observed in comparing the chromatogram of a
petroleum ether extract of coal tar pitch (Fig. 4) with a similar
(4)
analysis found in reference [12].
3.2. Qualitative results
3.2.1. Identification by RRT
In this paper, all the samples in Table 1 were analyzed in this
way. As this method is one of the methods widely used in the
identification of components in chromatographic analysis, it is
not necessary to list the detailed data of RRT here.
3.2.2. Identification by stop-flow UV scanning
The identification of fluoranthene and pyrene in petroleum
(5)
ether extract of coal tar pitch by this method is shown as exam-
ples in Fig. 5A and B, respectively. It can be seen that the
ultraviolet spectra of components in samples are exactly identi-
cal with those of corresponding pure reagents.
This method is reliable, but requires a larger injection size (for
example, 15 ␮l or 20 ␮l), and only one component is identified
in each run.
3.2.3. Identification by fluoresbene characteristic Ф index
Typical examples of this method are shown in Figs. 1–3. In
Fig. 2, the UV wavelength is 254 nm, fluorescence excitation
wavelength (Ex ) is 380 nm and emission wavelength (Em ) is
460 nm.
At wavelengths of “Ex = 265 nm, Em = 370 nm”;
“Ex = 250 nm, Em = 380 nm”; “Ex = 336 nm, Em = 374 nm”;
176 C. Zhang et al. / J. Chromatogr. A 1167 (2007) 171–177

Table 5
Response factors of PAHs determined by different methods
No Component Absorbance method HPLC method Average Deviation (%)

1 Naphthalene 0.089 0.088 0.088 0.569

2 Acenaphthylene 0.062 0.055 0.056 6.779
3 Biphenyl 0.452 0.437 0.445 1.798
4 ␤-Methylnaphthalene 0.074 0.076 0.075 1.333
5 Acenaphthene 0.034 0.038 0.036 5.555
6 Fluorene 0.259 0.267 0.263 1.521
7 2,6-Dimethylnaphthalene 0.082 0.078 0.080 2.500
8 2,3-Dimethylnaphthalene 0.069 0.063 0.066 4.545
9 Phenanthrene 1.000 1.000 1.000 0.000
10 Anthracene 2.024 2.126 2.075 2.272
11 Fluoranthene 0.310 0.282 0.291 3.093
12 Pyrene 0.226 0.213 0.220 3.182
13 9-Methylanthracene 2.653 2.741 2.697 1.631
14 Chrysene 0.930 0.952 0.934 1.927
15 Benz[a]anthracene 0.723 0.747 0.735 1.633
16 Benzo[e]pyrene 0.354 0.365 0.360 1.389
17 Perylene 0.154 0.162 0.158 2.532
18 Benzo[a]pyrene 0.385 0.379 0.382 0.785
19 7,12-Dimethylbenz[a]anthracene 0.477 0.483 0.480 0.625
20 1,3,5-Triphenylbenzene 0.490 0.486 0.488 0.409
21 2,3-Benzfluorene 0.271 0.255 0.263 3.042
22 Triphenylene 1.840 1.876 1.858 0.969
23 Dibenz[ah]anthracene 0.080 0.081 0.081 0.00
24 Benz[j]fluoranthene 0.310 0.320 0.315 1.587
25 Benz[k]fluoranthene 0.333 0.327 0.330 0.909
26 m-Terphenyl 0.135 0.141 0.138 2.174
27 p-Terphenyl 0.134 0.124 0.129 3.876
28 o-Terphenyl 0.142 0.138 0.140 1.429
29 Benzo[ghi]perylene 0.231 0.197 0.214 7.944
30 Coronene 0.795 0.807 0.802 0.623
Average 2.221

and “Ex = 250 nm, Em = 315 nm”; components having high
responses are chrysene, anthracene, pyrene and benzene
homologues, respectively.
By changing Ex and Em and comparing the Ф indexes with
those of pure reagents, each of the components can be effectively
identified.
Typical comparisons of Ф indexes of PAHs in residual oil
from ethylene production and thermal cracking residual oil are
listed in Tables 2 and 3, respectively. From the tables, it can
be seen that different PAHs have different Ф indexes, and the
method shows a high selection specificity. The Ф indexes of
benzo[a]pyrene, chrysene, perylene, and coronene are all very
large. The Ф indexes of the components are very close to those
of the pure reagents. Therefore, the identification of the peaks
is reliable.
3.2.4. UV characteristic index V method
Typical results with this method are shown in Fig. 4 and listed
in Table 4 for a petroleum ether extract of coal tar pitch.
From Fig. 4, it is clear that the response of a PAH is
greatly different at different UV wavelengths. From Table 4,
for most PAHs, the ratio of responses at different wave-
lengths in pure state is fairly close to that in practical
samples.
Compared with UV scanning method, the V method is supe-
rior in that it does not require high component concentrations
and saves much labor.
3.2.5. Comparison of different qualitative methods
Based on the data presented above, each method has its advan-
tages and disadvantages. The multi wavelength UV method
supplements the stop-flow UV scanning method which is reli-
able but laborious. For the analysis of complex samples from
coal and petroleum, a combination of different methods is more
desirable.
3.3. Quantitative results
3.3.1. Response factors (F) of PAHs
Response factors of 30 PAHs determined by the HPLC
peak area method and/or UV absorbance method are listed in
Table 5. From this table it can be seen that the deviation between
HPLC method and UV absorbance method was small. The UV
absorbance method was quick and convenient.
3.3.2. Quantitative results of typical practical samples
The quantitative analyses results of “Daqing” thermal crack-
ing residual oil by the E-F method are listed in Table 6. From
 C. Zhang et al. / J. Chromatogr. A 1167 (2007) 171–177 177

Table 6 2. A valuable external standard and response factor method

Quantitative analysis of “Daqing” thermal cracking residual oil (wt.%) (E-F method) to quantitatively determine PAHs was pre-
Components E-F 1 E-F 2 Average Deviation (%) sented and 30 F values were supplied for the first time by the
Naphthalene 0.130 0.156 0.143 0.013
authors.
Biphenyl 0.082 0.078 0.080 0.002 3. As an extensive and systematic study of the analysis of PAHs,
␤-Methylnaphthalene 0.181 0.174 0.178 0.004 this work will be a useful reference for the analysis of PAH
2,3-Dimethylnaphthalene 0.217 0.209 0.213 0.004 mixtures.
Acenaphthene 0.539 0.504 0.523 0.019
Phenanthrene 0.109 0.099 0.102 0.003
Anthracene 0.044 0.042 0.043 0.001
Acknowledgments
2,6-Dimethylnaphthalene 0.220 0.212 0.216 0.004
p-Terphenyl 0.018 0.016 0.017 0.001 The authors gratefully acknowledge the financial support
Fluoranthene 0.306 0.291 0.299 0.008 from the Natural Science Foundation of China (90210034), and
Pyrene 1.134 1.089 1.112 0.023 the National High-Tech Research and Development Program
m-Terphenyl 0.101 0.099 0.101 0.001
2,3-Benzfluorene 0.088 0.075 0.082 0.007
(Program, 2004CB217602).
Chrysene 0.126 0.172 0.149 0.023
Benz[j]fluoranthene 0.771 0.673 0.722 0.049 References
Benzo[e]pyrene 0.583 0.562 0.573 0.011
Perylene 0.688 0.711 0.699 0.012 [1] U.H. Yim, S.H. Hong, W.J. Shim, Chemosphere 68 (2007) 85.
Benz[k]fluoranthene 0.596 0.666 0.631 0.035 [2] V. Flotron, J. Houessou, A. Bosio, C. Delteil, A. Bermond, V. Came, J.
Benzo[a]pyrene 0.299 0.287 0.293 0.006 Chromatogr. A 999 (2003) 175.
7,12-Dimethylbenz[a]anthracene 0.128 0.110 0.119 0.009 [3] T. Ohura, T. Amaga, T. Sugiyama, M. Fusaya, H. Matsushita, Atmos.
1,3,5-Triphenylbenzene 0.077 0.074 0.075 0.001 Environ. 38 (2004) 20.
Coronene 2.679 2.663 2.671 0.008 [4] V. Pino, J.H. Ayala, A.M. Afonso, V. González, Anal. Chim. Acta 477
Average 0.011 (2003) 81.
[5] R.A. Gimeno, E. Comas, R.M. Marce, J. Ferre, F.X. Rius, F. Borrull, Anal.
Chim. Acta 498 (2003) 47.
[6] X. Yan, Y. Wu, Y. Sun, J. Chromatogr. A 724 (1996) 337.
this table it can be seen that the deviation% between duplicates [7] N. Kishikawa, M. Wada, N. Kuroda, S. Akiyana, K. Nakashima, J. Chro-
is from 0.001 to 0.050%. These results indicate that the E-F matogr. B 789 (2002) 257.
method is suitable to be used for complex heavy liquid samples. [8] J.F. Schabron, R.J. Hurtublse, Anal. Chem. 49 (1977) 2253.
Although some error still exists, it can handle multi-component [9] W.M. Coleman, D.L. Wooton, H.C. Dorn, L.T. Taylor, J. Chromatogr. 23
heavy samples more conveniently than other methods. (1976) 419.
[10] A.J. Packham, R. Fielden, J. Chromatogr. 552 (1991) 575.
[11] S.A. Wise, S.N. Chesler, H.S. Hertz, L.R. Hilpert, W.E. May, Anal. Chem.
4. Conclusion 49 (1977) 2306.
[12] D.W. Grant, R.B. Meiris, J. Chromatogr. 142 (1977) 339.
1. Four qualitative methods have been examined systemati- [13] T. Nielsen, J. Chromatogr. 150 (1979) 147.
cally to determine PAHs in products derived from coal and [14] C.M. Zhang, J.N. Zheng, Chin. Fuel Chem. Processes 31 (2000) 81.
petroleum.