Trends Trends in Analytical Chemistry, Vol. 30, No.

11, 2011

A review of techniques for

the determination of polycyclic
aromatic hydrocarbons in air
Sudhir Kumar Pandey, Ki-Hyun Kim, Richard J.C. Brown

We provide an extensive review of the common methodologies employed in the analysis of airborne polycyclic aromatic
hydrocarbons (PAHs). The review focuses on gas-chromatography-based approaches, in the light of their universal application
with excellent separation, resolution, and sensitivity.
We first describe collection methods for airborne PAHs in the gas and particle phases. We then evaluate the efficiency of
extraction techniques employed for separating target PAHs from sampling media, using conventional solvent-based and emerging
thermal-desorption approaches.
We also describe commonly employed analytical methods with respect to their applicability to PAHs in gas and particle
phases, collected from diverse environmental settings. As an essential part of basic quality assurance, we examine each method
with special emphasis on key parameters (e.g., limit of detection and reproducibility).
Finally, we address the likely directions of methodological developments, their limitations, and the future prospects for PAH
analysis.
ª 2011 Elsevier Ltd. All rights reserved.

Keywords: Airborne polycyclic aromatic hydrocarbon; Analytical discrepancy; Analytical methodology; Extraction; Gas chromatography; Limit of
detection (LOD); Polycyclic aromatic hydrocarbon (PAH); Reproducibility; Solvent-free; Thermal desorption

1. Introduction acronyms). In Europe, ambient air legis-

Sudhir Kumar Pandey+,
Ki-Hyun Kim*
Atmospheric Environment
Laboratory, Dept. of
Environment & Energy, Sejong
University, 98 Goon Ja Dong,
Gwang Jin Goo, Seoul 143-747,
Republic of Korea
Richard J.C. Brown
Analytical Science Division,
National Physical Laboratory,
Hampton Road,
Teddington TW11 0LW, UK
*
Corresponding author.
Tel.: +82 2 3408 3233;
Fax: +82 2 3408 4320;
E-mail: khkim@sejong.ac.kr
+
Presently at: Dept. of Botany,
Guru Ghasidas Central Univer-
sity, Bilaspur (C.G.), 495009,
India
Polycyclic aromatic hydrocarbons (PAHs)
are a group of over 100 different chemi-
cals that are known to be formed typically
during incomplete combustion of organic
matter at high temperatures [1]. Their
major sources in the atmosphere include
industrial processes, vehicle exhausts,
waste incineration, and domestic heating
emissions, while they can also be released
naturally {e.g., forest fires [2]}. They are
ubiquitous in the environment and con-
tain two or more fused benzene rings in
linear, angular or cluster arrangements.
In addition to the frequency with which
they occur in the environment, proof of
their mutagenicity and carcinogenicity led
to some of them being selected as priority
pollutants (e.g., 16 PAHs) by the US
Environmental Protection Agency (EPA)
(Table 1). The World Health Organization
(WHO) added 17 additional PAHs to make
a total of 33 PAHs under its regulation
[3]. (In Table 1, these 33 compounds are
listed with their three-capital-letter
lation targets benzo[a]pyrene (with an
annual target value of 1 ng/m3) because
this compound carries the highest toxic
load (defined as concentration multiplied
by toxicity) of any airborne PAH.
Because of their thermally stable struc-
ture, PAHs generally exhibit a high melt-
ing point, a high boiling point, and a low
vapor pressure. In the atmosphere, they
are distributed between gas and particle-
bound phases [4–9]. This phase parti-
tioning is largely regulated by changes in
atmospheric conditions (e.g., temperature
and relative humidity), and the physical
properties of the PAHs themselves [10–
13]. Consequently, lighter PAHs tend to be
preferentially enriched in the gas phase,
while the heavier ones show almost
complete association with particles [14] –
indeed, in Europe, air-quality legislation to
limit PAH concentrations in air sets target
values for only particulate-bound PAHs
(benzo[a]pyrene in the PM10 particulate
phase). For these reasons, it is a challenge
to acquire an accurate profile of these

1716 0165-9936/$ - see front matter ª 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.trac.2011.06.017
Trends in Analytical Chemistry, Vol. 30, No. 11, 2011 Trends

Table 1. Brief description of the 33 most common target polycyclic aromatic hydrocarbons (PAHs) (16 + 17 compounds)

Order PAHs Abbreviation Chemical CAS number Number of fused Molar mass (g/mol) Structure
formula benzene rings
[A] 16 priority PAHs designated by the US Environmental Protection Agency (EPA): most commonly referred in this review
1 Naphthalene NAP C10H8 91-20-3 2 128

2 Acenaphthylene ACN C12H8 208-96-8 3 152

3 Acenaphthene ACL C12H10 83-32-9 3 154

4 Fluorene FLR C13H10 86-73-7 3 166

5 Anthracene ANT C14H10 120-12-7 3 178

6 Phenanthrene PHN C14H10 85-01-8 3 178

7 Fluoranthene FLT C16H10 206-44-0 4 202

8 Pyrene PYR C16H10 129-00-0 4 202

9 Benzo(a)anthracene BAA C18H12 56-55-3 4 228

10 Chrysene CHY C18H12 218-01-9 4 228

11 Benzo(a)pyrene BAP C20H12 50-32-8 5 252

12 Benzo(b)fluoranthene BBF C20H12 205-99-2 5 252

13 Benzo(k)fluoranthene BKF C20H12 207-08-9 5 252

14 Benzo(ghi)perylene BGP C22H12 191-24-2 6 276

15 Indeno[1,2,3-cd]pyrene ICP C22H12 193-39-5 6 276

16 Dibenz(a,h)anthracene DBA C22H14 53-70-3 6 278

[B] 17 PAHs based on their probable carcinogenic and mutagenic behavior by the World Health Organization (WHO)

17 1-Methylphenanthrene MPT C15H12 832-69-9 3 192

18 Benzo[a]fluorene BAF C17H12 238-84-6 4 216

http://www.elsevier.com/locate/trac 1717
Trends Trends in Analytical Chemistry, Vol. 30, No. 11, 2011

Table 1. (continued)

Order PAHs Abbreviation Chemical CAS number Number of fused Molar mass (g/mol) Structure
formula benzene rings
19 Benzo[b]fluorene BFR C17H12 243-17-4 4 216

20 Benzo[ghi]fluoranthene BGF C18H20 203-12-3 5 226

21 Cyclopenta[cd]pyrene CCP C18H10 27208-37-3 5 226

22 Benzo[c]phenanthrene BCP C18H22 195-19-7 4 228

23 Triphenylene TPN C18H12 217-59-4 4 228

24 5-Methylchrysene MCS C19H14 3697-24-3 4 242

25 Benzo[j]fluoranthene BJF C20H12 205-82-3 5 252

26 Benzo[e]pyrene BEP C20H12 192-97-2 5 252

27 Perylene PRL C20H12 198-55-0 5 252

28 Anthanthrene ATT C22H12 191-26-4 6 276

29 Coronene COR C24H14 191-07-1 6 300

30 Dibenzo[a,e]pyrene DAP C24H14 192-65-4 6 302

31 Dibenzo[a,h]pyrene DHP C24H14 189-64-0 6 302

32 Dibenzo[a,i]pyrene DIP C24H14 189-55-9 6 302

33 Dibenzo[a,l]pyrene DLP C24H14 191-30-0 6 302

harmful substances in environmental matrices, includ-
ing the atmosphere.
For the analysis of PAHs in air, gas chromatography
(GC) has been the most common analytical separation
mechanism in combination with a number of sampling
approaches and sample-preparation steps. The prefer-
ence for GC over other systems [e.g., liquid chromatog-
raphy (LC)] has been demonstrated by such factors as its
greater selectivity, resolution, and sensitivity [3]. To
improve selectivity and to ensure the robustness of

1718 http://www.elsevier.com/locate/trac
Trends in Analytical Chemistry, Vol. 30, No. 11, 2011 Trends

sample identification and quantification, GC is often used Table 2. Common collection methods (media) for the analysis of air-
in combination with mass spectrometry (MS) to produce borne polycyclic aromatic hydrocarbons (PAHs) in particulate phase
very powerful GC-MS coupled techniques, which are able
Collection/sorbent material Target PAHs Ref.
to yield even lower limits of detection (LODs) than GC
modea
alone.
This review offers a critical evaluation of the state- [A] Solvent-based extraction
(i) Glass fiber filter (GFF)
of-the-art information on the methodological choices
A [29]
available for, and recent developments in, PAH analysis B [30,31,34,58]
in air. We discuss the procedures involved in sampling C [25,32,54,55]
and/or preconcentration, sample preparation, and final E [33]
determination (or detection) of PAHs in air, in order to
assess and to review the nature of GC-based applications (ii) Quartz fiber filter (QFF)
B [37]
for their analysis. To this end, we initially discuss con- C [36,39]
ventional solvent-based sample collection along with D [38]b
standard preparation methodologies. Moreover, we also E [57]
evaluate simplified solvent-free methods, as alternatives
to solvent-based methods, and compare for their appli- (iii) Teflon filters
C [17,40]
cability in this area. Finally, we give a detailed descrip-
tion of most of the essential components involved in the (iv) Mixed filters
basic quality-assurance procedures required for the XAD-4 coated glass denuder B [47]
determination of PAHs in air. Teflon-coated glass fiber filter E [35]

[B] Solvent free analysis (thermal desorption)

2. The collection of PAHs
For the collection of airborne PAHs in both vapor and
particle phases, large volumes of air must be sampled to
concentrate them on the sorbent material (vapor phase)
or a suitable filter material (particulate matter). This is
because their concentration in air at most locations is
relatively low (of the order of ng/m3). The most common
sorbent used for gas-phase enrichment is polyurethane
foam (PUF). However, other sorbent materials (e.g.,
XAD-2, XAD-4, Carbopack C, and Tenax) have also been
employed occasionally [15–18] (Table 2). High-volume
samplers are operated at sufficiently high flow rates (34–
1250 L/min) and for a long enough to guarantee suffi-
cient sample volume for quantitative analysis (e.g., 150–
900 m3 for gas phase and sometimes up to 3000 m3 for
particle phase).
Gas-phase samplers are usually protected from any
particulate-phase material in the air sampled by the
addition of a pre-filter at the start of the sampling train.
In some situations where the aim of the study is analysis
of both gaseous and particulate phases, this pre-filter
may then be used for analysis of PAHs in the particulate
phase, although the size-fraction characteristics of the
collected particulate material will generally be unknown
and unregulated (unless a size-selective inlet is em-
ployed). In addition, gas-phase PAHs may be sampled
using traditional PM10 air samplers, where the size-
fraction characteristics of the particulate phase sample
are well known. In this case, the PUF cartridge sits in a
tube beneath the filter that collects the PM10 particulate
material. This is currently done at a limited number of
monitoring sites as part of the UK PAH air quality
(i) QFF
A [63]
B [23,59,61]
(ii) Teflon filter B [20]
a
Describes the combination of target PAHs commonly selected in
many previous studies (A = some of 16 priority PAHs of US EPA,
B = some of 16 + some extra, C = 16 PAHs only, D = 16
PAHs + some extra, and E = All PAHs except 16 PAHs).
b
Sampling with personal pumps.
monitoring network [19] (which assists the UK to com-
ply with the European Fourth Air Quality Daughter
Directive, that requires Member States to monitor for
B[a]P and six other PAHs). Upon completion of air
sampling, these sorbents and filter materials are treated
with organic solvents (e.g., dichloromethane, toluene,
hexane, acetonitrile, and methanol) either individually
or in combination to extract target PAHs. Extraction can
be performed by manual washing or via automated set-
ups (e.g., Soxhlet extraction or accelerated solvent
extraction). It is important to extract the PAHs from the
sorbent material using as little solvent as possible in
order to maximize the mass fraction of analyte in the
solvent, thereby improving the overall method LOD. The
final determination of PAHs is then made, preferably
with GC-MS.
Many studies aiming to monitor PAHs in air have
attempted to measure both particle and gas-phase PAHs
simultaneously (Figs. 1 and 2). There are also a number
of studies that have focused only on either one of the
two, while many studies show a preference for the par-
ticulate phase [20–26]. As the distribution of airborne

http://www.elsevier.com/locate/trac 1719
1720

Trends
http://www.elsevier.com/locate/trac

Detection MS FID ECD

(B) Some of 16 (C) 16 priority (D) 16 (E) All extra (C) 16 priority (E) All extra
Target PAHs (A) Some of
PAHs + extra PAHs only priority PAHs PAHs only
16 PAHs
+extra

GFF+ Teflon QFF + GFF +

Urban area GFF + solvent
solvent coated solvent solvent
desorption [29 ], QFF [36, 39], GFF [31,
based GFF desorption desorption
QFF + TD [59] PTFE [17], 32, and 58]
desorption[ [35], [39] [33]
GFF [25] + + solvent GFF +
30, 58] desorption
solvent solvent
Teflon &
desorption desorpti
QFF +TD
[23, 64] on [33]

Industrial GFF + solvent

area GFF + QFF +
desorption [29]
solvent solvent
desorption

Trends in Analytical Chemistry, Vol. 30, No. 11, 2011

desorption
[55] [38]
Rural area GFF + solvent QFF +
desorption [29] solvent
desorpti
on [57]
Teflon +
Indoor area solvent
desorption
[40]

QFF +
Lab and solvent
others desorption
[37]

Figure 1. Plot of all possible combinations of detection and sampling methods for polycyclic aromatic hydrocarbons (PAHs) in the particle phase.
 Trends in Analytical Chemistry, Vol. 30, No. 11, 2011
Detection MS FID

(C)16 priority
Target PAHs (A) Some of 16 (B) Some of 16 (C) 16 priority (D) 16 priority (E) All extra
PAHs only
PAHs PAHs + extra PAHs only PAHs +extra

Urban area Carbopack [15] and PUF and XAD and PUF + PUF/XAD
PUF [29] + solvent XAD + PUF + -2/PUF +
PUF + solvent
desorption solvent solvent solvent
solvent desorption
desorption desorption desorption
desorption
[47, 58] [ 31, 32] [25] [48]
[17, 36]

Industrial PUF + solvent

area PUF +
desorption [29] solvent
desorption
[55]

Rural area
PUF +
solvent
desorption
[57]
PUF +
http://www.elsevier.com/locate/trac

Indoor area XAD + solvent

solvent
desorption [18] &
desorption
GC-columns +TD
[40]
[50]

Lab and PDMS traps +TD PDMS alone &

others [62] Mixed (PDMS
+Tenax bed) +
TD [49, 68]

Figure 2. Plot of all possible combinations between the detection and sampling methods for polycyclic aromatic hydrocarbons (PAHs) in the

Trends
1721

vapor phase.
Trends Trends in Analytical Chemistry, Vol. 30, No. 11, 2011

PAH basically occurs in both gas and particle-bound
phases, all the sample-collection methods are designed to
cover (1) each of the two fractions individually or (2)
their sum fractions at the same time [27,28].
2.1. Collection of particle-phase PAHs
Because of complexities involved in the analysis of PAHs,
some studies were simply confined to the particulate
fractions to explore the distribution behavior of PAHs.
For the collection of both particle-bound and particle
plus gaseous PAHs in air, glass-fiber filter (GFF) and
quartz-fiber filter (QFF) were used commonly (Table 3).
However, on rare occasions, some other materials (e.g.,
Teflon) were also employed individually or in combina-
tion (Table 3). When particulate-phase PAHs are
sampled, it is usually the PM10 particulate phase of
ambient air that is considered, since this is currently the
size fraction most relevant to human-health studies and
to regulation. Note that filter materials often used to
collect particulate samples for metals analysis (e.g., cel-
lulose membranes) rarely have sufficient strength or
porosity to sustain the very high flow rates used for
sampling particulate-phase PAHs without damage or
filter clogging.
2.1.1. Glass-fiber filter (GFF). GFFs are the most fre-
quently employed collection devices for particle-phase
PAHs. The target PAHs mainly include 16 priority PAHs
and their derivatives but are occasionally extended to
additional PAHs; their sampling by GFFs has been made
from diverse locations from rural to industrial areas.
Although the total sample volume and flow rate set for
their collection differ between studies, most have passed
sufficiently large volumes of air through these filter
materials to ensure enrichment (for sufficient extraction)
with the common analytical methods.
Gustafson and Dickhut [29] made measurements of
PAHs in air at four sites representative of rural,

Table 3. Common collection methods (media) for vapor-phase polycyclic aromatic hydrocarbons (PAHs) in air

Collection/sorbent material Target PAHs modea Ref.

[A] Solvent based
(i) Polyurethane (PUF)
B [58]
C [29,36,40,55]b
D [31,32]
E [25,57]

(ii) XAD
XAD-4 coated glass denuder B [47]
XAD-2 C [17]

(iii) Combination of PUF and XAD

PUF/XAD-2/PUF C [48]

(IV) Others
Carbopack C (30 mesh, 400 mg)c A [15]
Glass wool filter, condenser, glass fiber filter, Teflon filter, and a Tenax trapd C [16]

[B] Solvent free (thermal desorber)

(i) XAD
XAD-2 A [18]

(ii) Polydimethylsiloxane (PDMS)

Quartz multi-channel PDMS traps A [62]
PDMS traps C [68]

(iii) Mixed sorbents

Mixed sorbent beds ((PDMS foam + PDMs particle + Tenax bed) C [49]

(IV) Others
Fan-Loy sampler (SPB-5 GC columns (0.75-mm i.d. and 7-lm film thickness) A [50]
a
Describes the combination of target PAHs commonly selected in many previous studies (A = some of 16 priority PAHs of US EPA, B = some of
16 + some extra, C = 16 PAHs only, D = 16 PAHs + some extra, and E = All the PAHs except 16 PAHs).
b
Personal monitors.
c
Diffusion-based sampling.
d
On-line.

1722 http://www.elsevier.com/locate/trac
Trends in Analytical Chemistry, Vol. 30, No. 11, 2011
semi-urban, urban, and industrialized areas (southern
Chesapeake Bay, USA). These authors actively passed
172–365 m3 of air (flow rate = 510–750 L/min) though
GFFs to collect particle-phase PAHs. They were able to
characterize 14 out of 16 priority PAHs over a wide
concentration range (pg–ng/m3).
Omar et al. [30] also collected 16 priority PAHs in
airborne PM10 on GFFs from urban roadside and rural
location (Kuala Lumpur, Malaysia). However, they used
comparatively large sample volumes [i.e. 1612–
3224 m3 (at flow rate of 1120 L/min)] to ensure their
quantification at sub-ng/m3 level.
In another report made by Mandalakis et al. [31], 20
PAHs (16 priority PAHs + Me-phenanthrenes, dimethyl-
phenanthrenes, Me-chrysene, and COR) were collected
on GFFs from diverse locations (an urban center, a
background site, and the adjacent coastal area). These
authors collected 450–900-m3 air samples to ensure the
enrichment of the particle-bound PAHs on GFFs.
In a similar study, Tsapakis and Stephanou [32] col-
lected 24 (16 priority + 8 others) particle-bound PAHs
by passing 900 m3 of urban air (flow rate = 600 L/min)
from an urban area in Heraklion, Greece, through GFFs.
GFFs have also been employed to collect a number of
PAH derivatives [e.g., oxy-PAHs and nitro-PAHs (often
formed by PAH degradation – see later)], apart from 16
priority PAHs. Castells et al. [33] used GFFs as the col-
lection media for seven oxy-PAHs and nitro-PAHs pres-
ent in urban-aerosol samples in Spain by using a very
large sampling volume of 1440 m3 (flow rate = 1000 L/
min). In another study, diverse PAH components in PM
(25 PAHs + 12 nitro-PAHs + four oxy-PAHs) were
quantified via GFF sampling from a heavily trafficked
square in Basel, Switzerland (sample volume = 960 m3
at a flow rate of nearly 666 L/min) [34].
On rare occasions, modifications of GFFs have also
been used for the collection of particle-bound PAHs.
Dimashki et al. [35] used Teflon-coated GFFs (often re-
ferred to as Emfab) to collect particle-phase PAHs from
urban atmospheres in Birmingham, UK, and Damascus,
Syria. The total volume of air passed through these
collection devices was 900–1000 m3 (flow rate = 650–
750 L/min).
2.1.2. Quartz-fiber filter (QFF). QFF is another preferred
medium for the collection of particle-bound PAHs. Park
et al. [36] collected 16 priority PAHs on QFFs (102 mm
diameter) in an urban area of Seoul (Korea) using an
active sampling method. Lottmann et al. [37] collected
particulate PAHs (eight priority PAHs + COR) on QFFs
(150 mm diameter) by drawing a total volume of
720 m3 (1250 L/min) at a university campus
(Strasbourg, France).
QFFs have also been employed to collect particle-phase
PAHs with relatively low sampling volumes. Chaspoul
et al. [38] attempted to collect 16 priority PAHs and six
Trends
nitro-PAHs simultaneously from industrial workplaces
at a flow rate of 1 L/min on QFFs (total sample volume
0.5 m3). Chantara and Sangchan [39] used QFFs with
a mini-volume air sampler to collect PAHs in PM10 from
diverse urban areas. Note that QFFs can also be suitable
as collection media for solvent-free methods [e.g., ther-
mal desorption (TD)] {e.g., PAHs (many of the 16 pri-
ority PAHs) and oxy-PAHs in PM2.5 were collected on
QFFs (sample volume = 24 m3 at a sampling flow rate of
1000 L/min) from an urban area of Germany (Augsburg
City) prior to thermal desorption GC-MS analysis [23]}.
2.1.3. Teflon filters and other media. Although not as
common as other fiber-type filters, Teflon filters (e.g., 37-
mm diameter) are often reported for particulate-PAH
collection. As one example, these filters were mounted in
personal environmental monitors for the collection of
PAHs in PM2.5 (sampling duration of 8 h). Their sam-
pling was made from indoor locations divided into
smoking and non-smoking zones [40]. The flow rate of
these personal sampling pumps was 4 L/min (with a
variability of 5%).
In the light of its high thermal conductivity, alumi-
num foil has also been tested to sample particle-phase
PAHs along with other semi-volatile organic compounds
[41]. The authors collected urban PM10 samples on an
aluminum-foil substrate installed in an Andersen eight-
stage cascade impactor at a flow rate of 28.3 L/min (50-
h duration). Prior to use, the aluminum foils were heated
to 500C for 24 h. The authors stated that aluminum foil
has very high efficiency for thermal desorption, while
QFF can suffer from pyrolysis of long-chain compounds
during the thermal-desorption process.
2.1.4. Degradation of PAHs during sampling. It has been
known for some time that, once particulate-bound PAHs
become immobilized on air filters following sampling,
they become much more susceptible to degradation from
oxidizing compounds, particularly ozone, but also NO2,
present in the sampled air, which are drawn through the
collected PM. PAHs can be oxidized upon the reaction
with ozone and NO2 to produce oxy-PAHs and nitro-
PAHs, respectively [42]. Such post-sampling reactions
have been shown by Schauer et al. [43] to result in
conversion of up to 50% of the sampled PAHs. This, in
turn, may result in a substantial underestimation of the
concentrations of PAHs present in ambient air. In order
to determine the quantity of degradation present in the
sampled PM, the analysis of oxy-PAHs and nitro-oxy-
PAHs must be undertaken which is considered more
challenging [34]. In addition, there is expected to be a
small quantity of these compounds present in ambient
air as a result of gas-phase reactions between particu-
late-bound PAHs and oxidizing compounds in air. The
degradation of PAHs on sampled filters is one of the main
reasons that it is prudent to limit the length of sampling
http://www.elsevier.com/locate/trac 1723
Trends
periods for each PAH sample so as to minimize the effect
of oxidizing gases. Indeed, the European Fourth Daugh-
ter Directive [44] mandates Member States to use sam-
pling periods of 24 h. Whilst the effect of degradation
may be noticeable on short-term monitoring studies, it is
thought that the effect on annual average values (e.g.,
those produced by fixed air-quality networks) is likely to
be small. As the annual variations in the concentration
of PAHs and ozone are in anti-phase, there is very little
effect on measured PAH concentrations. For example, in
the winter (e.g., increased fuel use), PAH concentrations
become high whilst ozone concentration is low. Con-
versely, in the summer when the ozone is high, PAH
concentrations are low. Hence, although relative deg-
radation rates can rise, these have little effect on the
annual average. Moreover, judicious timing of the
automatic filter-changeover facility on many of the
current air-quality-monitoring samples can allow filter
changes to occur at dawn, to take advantage of the
diurnal variation in ozone concentrations and to achieve
a large period of time during the night prior to sample
changeover when ozone is low. Thus, the degradation to
PAHs collected during the day, and subsequently over-
night, is minimized. A full understanding of the role of
ozone in PAH degradation is still being developed [45].
Recent developments to solve the ozone-degradation
problem have included use of an ozone denuder in
combination with the PM10 sampler [46]. This innova-
tion removes ozone from the sampled air and prevents
degradation of PAHs within the sampled PM. However,
these devices may be less effective in humid and wet
conditions, so the same group is currently working on an
internal heater for the ozone scrubber to solve this
problem.
2.2. Collection of gas-phase PAHs
If we review the literature available for airborne-PAH
analysis, the dominant portion is found to cover 16
target PAHs (of the US EPA) in both particulate and
vapor phases. As the collection devices for particle-phase
PAHs have been aforementioned, we now focus on the
substances used for the collection of gas-phase PAHs in
the following sub-sections. Because of the low concen-
trations of PAHs present in air, their collection usually
relies on pumped sampling (where a known volume of
air is drawn through the sorbent) rather than diffusive
sampling (where vapor is allowed to diffuse into the
sorbent). This is because diffusive samplers are in many
times not simple to calibrate for quantitative analysis.
2.2.1. Polyurethane foam (PUF). PUF has been utilized
as the most common means of enriching gas-phase
PAHs. It is applicable for a wide range of PAHs (16
priority PAHs and many others, especially their nitro
and oxy derivatives). Gustafson and Dickhut [29] col-
lected gas-phase PAHs (14 out of 16 priority PAHs) on
1724 http://www.elsevier.com/locate/trac
Trends in Analytical Chemistry, Vol. 30, No. 11, 2011
two PUF plugs by passing 172–365 m3 of air (flow
rate = 510–750 L/min) at both rural and industrial
sites. Mandalakis et al. [31] used a PUF plug
(length = 8.0 cm, diameter = 7.5 cm) to enrich 20 PAHs
(16 priority PAHs + four others) by passing 450 m3 and
900 m3 of air from diverse locations (an urban center, a
background site, and the adjacent coastal area). In a
similar study, Tsapakis and Stephanou [32] collected 24
airborne PAHs (16 priority PAHs + eight others) by
passing 900 m3 of urban air (flow rate = 600 L/min)
through PUF.
The applicability of PUF samplers has also been dem-
onstrated in the determination of diverse PAH deriva-
tives (e.g., nitro-PAHs and oxy-PAHs) in the vapor
phase. Dimashki et al. [35] used PUF (80 mm diameter,
75 mm depth, density = 0.02 g/cm3) to collect vapor-
phase PAHs; they used a sampling volume of 900–
1000 m3 at a flow rate of 650–750 L/min from two
urban locations (Birmingham, UK, and Damascus,
Syria).
2.2.2. Collection on XAD. As an alternative to PUF,
Possanzini et al. [47] tested an XAD-4 (20–60 mesh)
coated-glass denuder (with QFF for particle) for the col-
lection of gas-phase PAHs. They assessed collection effi-
ciency (and capacity) over 6-h sampling at an air flow
rate of 6 L/min. The collection efficiencies of two-ring
and three-ring PAHs were found to exceed 90% [e.g.,
90% (PYR)–97% (NAP)].
2.2.3. Mixed sorbents/others. For the collection of gas-
phase PAHs, PUF samplers have been used most inten-
sively in combination with other sorbents (e.g., XAD-2).
Lee et al. [48] used a glass cartridge containing a 5-cm
PUF plug, followed by 3-cm XAD-2 resin, and finally a 2-
cm PUF plug to collect the gas-phase PAHs from an area
affected by traffic activities. Bertoni et al. [15] proposed a
molecular diffusion-based sampling device for the
determination of gaseous PAHs in air. These authors
used Carbopack C as the adsorbing material to enrich the
PAHs for the GC-MS analysis after two monthsÕ exposure
of the sorbent in urban or suburban air.
Wauters et al. [49] used a mixed sorbent bed {poly-
dimethylsiloxane (PDMS) foam (1 cm), PDMS particles
(120 mg), and a Tenax TA (60 mg)} for the collection of
16 priority PAHs by a personal air pump at a flow rate of
100 mL/min for 24 h. They compared their method to
the classical method relying on high-volume sampling
onto a PUF sampler for 24 h. The total volumes of
samples collected by these methods contrasted between
144 L (mixed bed) and 1296 m3 (classic method).
However, they found the new approach superior to the
conventional method, as the new approach yielded
much higher recoveries – ranging from 1.2 times better
for PHN to 35 times better for NAP – despite the much
smaller sampling volume.
Trends in Analytical Chemistry, Vol. 30, No. 11, 2011
An online sampling system was also designed to col-
lect the 16 priority PAHs emitted from a coal-fired flu-
idized bed combustor (sample volume = 60 L at a flow
rate of 2.5 L/min). This system was built to combine a
glass-wool filter (to remove particles), condenser (to
remove water), GFF, Teflon filter, and a Tenax trap (as
the trapping device for PAHs) [16]. In addition, for
assessing personal exposure to PAHs in urban-
community settings, Fan et al. [50] prepared a honey-
comb-like sampler with 320 sections of 1-cm long SPB-5
GC columns (0.75-mm i.d. and 7-lm film thickness) and
collected gaseous PAHs (NAP, ACN, ACL, FLR, PHN,
ANT, FLT, and PYR) on the inner surfaces of the col-
umns by molecular diffusion. Their method was suffi-
cient to enrich these gas-phase PAHs for further analysis
with TD-GC-MS.
3. Evaluation of recoveries for different extraction
techniques
As the analysis of PAHs generally relies on collection
media for pre-concentration, regardless of their phase,
one needs to know how effectively their extraction pro-
ceeds in the sampling stage. It has always been a chal-
lenge to extract all the target PAHs with sufficiently good
efficiency due to large variations in their physicochemi-
cal properties (e.g., volatility and reactivity). The use of
surrogate standards directly spiked to the sorbent or filter
material before sampling (or after sampling if the sam-
ples are measured Ôas receivedÕ and losses during sam-
pling are not considered) is usually recommended for
recovery calculations. For example, the best surrogate
standards are isotopically-labeled congeners (e.g., deu-
terated analogues of the PAHs being measured) [3]. It is
important to bear in mind that, whilst isotopically-
labeled PAH congeners are expected to mirror almost
exactly the physical and chemical properties of their
non-isotopically labeled counterparts, because they are
spiked onto collected PAHs, they can never exactly
replicate the matrix of the as-collected sample. This is
especially true for PAHs bound to particulate matter. The
recovery of isotopically labeled congeners may, therefore,
be slightly higher than that for the sampled PAHs –
although the inherent variability of recovery data makes
it very difficult to draw such conclusions with confidence.
Although many studies dealing with PAH determination
do not provide the full details of recovery data, numerous
efforts have been directed to the estimation of extraction
efficiencies [51,52]. The results of such efforts can be
examined for both the gas and particle phase.
3.1. Recoveries of solvent-based extraction for particle-
phase PAHs
After the collection of airborne particulate PAHs, the
filter samples are treated with a range of organic solvents
Trends
(e.g., hexane, dichloromethane, acetone and methanol)
individually or in combination. Moreover, to enhance
the recovery of PAHs from the filter materials, solvent-
based extraction is done by specialized apparatus (e.g.,
Soxhlet, microwave extraction, or ultrasonication) or
other techniques (e.g., pressurized fluid extraction)
[53,54].
Gustafson and Dickhut (28) extracted PAHs from GFF
samples by combining acetone, petroleum ether, and
dichloromethane as extractants in a Soxhlet apparatus.
The recoveries of surrogate PAHs (i.e. Anth-d10, BaA-
d12, and BaP-d12) were 92.6 ± 22.1%, 103.1 ± 9.5%,
and 79.3 ± 13.5%, respectively.
Omar et al. [30] estimated recovery of particle-phase
PAHs on GFFs with ultrasonication in dichloromethane
as 50.8–99.7%. However, these authors also reported
that the lighter PAHs (NAP, ACN, ACL, and FLR) could
not be accurately identified by this method due to low
recoveries and poor reproducibilities.
By contrast, Zhang et al. [55] extracted particle-phase
PAHs on GFF with hexane and cyclohexane and ob-
served superior recovery rates of 84.5% (NAP) to 108%
(BAP).
The European Committee for Standardization (CEN)
Working Group (TC264 WG21), which developed and
wrote the reference method for benzo[a]pyrene mea-
surement in air (the PAH subject to target-value
assessment in Europe), found relatively little difference in
performance between a number of solvents tested for
their extraction ability. Indeed, most solvents and sol-
vent combinations are allowed by the standard method,
as long as they achieve satisfactory recoveries when
analyzing NIST Standard Reference Material 1649a.
However, their standard does make the following rec-
ommendations for ‘‘preferred solvents’’, depending on
extraction technique [56]:
(1) Extraction under reflux: toluene
(2) Soxhlet extraction: toluene, dichloromethane, 1:1
hexane:acetone mixture
(3) Microwave extraction: 1:1 hexane:acetone mixture
(4) Accelerated solvent extraction: toluene, dichlorometh-
ane, 1:1 hexane:dichloromethane mixture
(5) Ultrasonic extraction: toluene, dichloromethane
As summarized in Table 2, many researchers have
attempted to describe the extraction efficiency of priority
PAHs. Extraction efficiencies may also be adversely af-
fected if solid-phase extraction (SPE) and/or sample
clean-up procedures are required prior to analysis. These
are regularly required for ambient air samples because of
frequent instances of carbonaceous and particulate
contamination after extraction. However, unlike those
common target PAHs, it has been an even greater
challenge to acquire optimum recoveries for their
derivatives (e.g., oxy-PAHs and nitro-PAHs) [34]. For
instance, nitro-PAHs and oxy-PAHs from GFF filters
were extracted in methanol/acetone with the assistance
http://www.elsevier.com/locate/trac 1725
Trends
of a microwave [34]. They were further purified with
SPE and analyzed by semi-preparative high-performance
liquid chromatography (HPLC). The GFF recovery rates
of nitro-PAHs and oxy-PAHs were in the range 80–90%.
Dimashki et al. [35] found relatively good recovery rates
of particle-bound PAHs on GFFs in the range 61–98%,
when extracted via Soxhlet with dichloromethane.
Albinet et al. [57] used QFFs for collecting nitro-PAHs
and oxy-PAHs in the particulate phase. They extracted
all the samples in dichloromethane and proceeded with
an LC purification procedure on an SPE cartridge. The
recovery rates of nitro-PAHs in the whole analytical
procedure were estimated as 14% (1-nitronaphthalene)
to 84% (7-nitrobenz[a]anthracene), while those of oxy-
PAHs were from 5% (1-naphthaldehyde) to 83%
(benzo[b]fluorenone).
It is usual practice in particulate-based PAH studies to
correct the final results based on the recovery of the
surrogate standards spiked into the real samples. Where
there is a direct isotopic analogue measurement avail-
able for the target PAH in question, this is simple. Where
there is no direct isotopic analogue measurement avail-
able for the target PAH in question, expert judgment
needs to be used to choose the most appropriate isoto-
pically-labeled PAH for performing the recovery correc-
tion. This is likely to be most chemically similar and
closely-eluting isotopically-labeled PAH from the gas
chromatogram.
3.2. Solvent-based recoveries of gas-phase PAHs
As for filter materials for collecting particulate PAHs,
sorbent materials for collecting gaseous PAHs can also
be subject to extraction by organic solvents with the
modification of various treatment techniques. Gustafson
and Dickhut [29] estimated recoveries of surrogate PAHs
(i.e. ANT-d10, BAA-d12, and BAP-d12) in PUF plugs.
These authors found recovery rates of 89.3 ± 18.5%,
100 ± 20.7%, and 87.2 ± 19.8%, respectively, based on
Soxhlet treatment with acetone, petroleum ether, and
dichloromethane (as extractants). Simcik et al. [58]
achieved recovery rates of 78–101% from surrogate
standards treated on PUF after extraction with dichlo-
romethane. Zhang et al. [55] collected gas-phase PAHs
on PUF plugs and then extracted them with hexane and
cyclohexane. They were able to yield recovery rates of
63.2% (Nap) to 108% (BaP). Supercritical fluid extrac-
tion (SFE) with dichloromethane (plus ultrasonication)
was employed to extract gas-phase PAHs from PUF [36].
The recovery rates of this application were 69–95%
(mean = 83%).
For the analysis of gas-phase nitro-PAHs (from PUF),
Dimashki et al. [35] employed Soxhlet apparatus in
dichloromethane and found recovery rates of 57–92%.
By contrast, the recovery of PAHs from Carbopack C
ranged from 53% (NAP) to 100% (CHY), when they
1726 http://www.elsevier.com/locate/trac
Trends in Analytical Chemistry, Vol. 30, No. 11, 2011
were extracted by a single solvent (toluene) at ambient
temperature [15].
According to the above comparison, the recovery rates
of gaseous PAHs vary over a wide range, depending
upon both the types of extraction technique and the
individual PAHs. However, compounds of low molecular
mass with relatively high volatility (e.g., NAP) tend to
show the lowest recoveries, regardless of phases (gas and
particle).
3.3. Extraction efficiency of solvent-free or thermal-
desorption methods
In recent years, many efforts were made to find an
alternative analytical procedure for PAHs. As most of the
previous analyses generally relied on the classic, solvent-
based approach, more recent studies were directed to-
ward use of solvent-free methods (e.g., TD systems).
Datasets on the extraction efficiency of these solvent-free
methods are still scarce. In most cases, the reliability of
such techniques cannot be assessed directly against
those of conventional solvent-based methods, regardless
of PAH-phase type.
3.3.1. Particle phase. Drooge et al. [59] evaluated
recovery ratios of some particle-bound PAHs collected
via QFFs based on comparative analysis between the
standardized liquid extraction (LE) method and the TD
method. In case of PHN and ANT, they found recovery
ratios in the TD method higher than those in the LE
method. However, the recovery of other PAHs in the TD
method was comparable to those in the LE method. Bates
et al. [60] were able to achieve their TD recoveries in the
range 95.5% (BGP) to 99.1% (BAA) for samples collected
on QFFs. In other research, Gil-Molto et al. [61] esti-
mated the recovery ratios of 12 PAHs (NIST SRM1649a
urban dust) on QFFs through comparative analysis by
the TD method. They found the recovery ratios of par-
ticulate PAHs in the range 95.01% (ICP) to 99.59%
(PHN).
3.3.2. Gas phase. In solvent-based methods, extractions
of gaseous PAHs have been conducted almost solely from
PUF. However, the solvent-free techniques (i.e. the TD
method) are sufficiently advanced to allow comparison
between diverse sorbent materials. Wei et al. [18] used a
microwave-assisted desorption (MAD) coupled to head-
space solid-phase microextraction (HS-SPME) as a one-
step sample-preparation method for PAHs with XAD-2
adsorbent. Although their method cannot be regarded as
fully solvent-free, they were able to exclude the use of
any toxic organic solvent. The eight target PAHs (out of
16 priority PAHs) on XAD-2 were then desorbed into
ethylene glycol and evaporated into the headspace via
microwave irradiation. These PAHs were then absorbed
directly on an SPME fiber in the headspace for final
Trends in Analytical Chemistry, Vol. 30, No. 11, 2011
detection by MS. Their study achieved a recovery rate of
>80% under optimized extraction conditions.
Fan et al. [50] estimated >85% recovery of eight
gaseous PAHs (NAP, ACN, ACL, FLR, PHN, ANT, FLT,
and PYR) from a honeycomb-like passive sampler (320
sections of 1-cm long SPB-5 GC columns) by a TD
method. However, large variabilities were observed in
the recoveries of FLT and PYR with relative standard
deviations (RSDs) of 36% and 30%, respectively. These
unstable trends were suspected to come from tempera-
ture inhomogeneity in the thermal-desorption system.
Wauters et al. [49] found the recovery of 16 priority
PAHs in the range of 80.5% (NAP) to 118% (BGP)
through combined application of a mixed sorbent-bed
and solvent-free TD-GC-MS analysis.
4. The solvent-based analysis of PAHs
Most of the previous studies aiming to investigate the
distribution of airborne PAHs (particle or gas phase)
have relied on solvent-based analysis, once the initial
collection of PAHs is completed on a suitable filter or
sorbent material. The PAHs bound on those media are
then extracted with organic solvents (in liquid phase) for
the final determination, preferably with GC-MS. These
conventional methods have been employed successfully
to describe the distribution pattern, source apportion-
ment, and phase distribution (gas or liquid) of PAHs
from diverse environmental settings (rural, urban,
industrial, indoor, workplace, and laboratory condi-
tions). In this section, we discuss the common and
modified approaches adopted in the solvent-based
determination of PAH concentrations, based on the key
phase-classification criteria (i.e. particle and gas), as we
did for sample-collection methods (Tables 4 and 5).
4.1. Solvent-based analysis of particle-phase PAHs
Gustafson and Dickhut [29] were able to analyze 14 out
of 16 priority PAHs using samples collected from diverse
locations around southern Chesapeake Bay, USA. These
authors conducted GC-MS analysis on particle-phase
PAHs collected on GFFs after extraction in dichloro-
methane with the aid of ultrasonication. The GC-MS-
based determination yielded PAHs in a concentration
range of 0.58 pg/m3 (DBA at a rural site) to 0.86 ng/m3
(CHY in an industrial area). Omar et al. [30] reported
BGP and COR as the most abundant PAHs in airborne
PM10 from urban roadside and rural location (Kuala
Lumpur, Malaysia) based on the GC-MS analysis after
solvent-based extraction of PAHs on GFFs.
Concentrations of 16 priority PAHs bound with PM10
samples (from diverse urban areas) were determined by a
GC-MS method after extraction in acetonitrile with an
ultrasonicator [39]. The MS was operated in single-ion
monitoring (SIM) mode to quantify characteristic ions of
Trends
selected PAHs with the aid of the deuterated internal
standard (closest to the analyte in molecular weight).
They were able to report the mean concentrations of
total PAH (sum of 16) in the range 2.7 ng/m3 (rural
area) to 16.6 ng/m3 (traffic area).
In another study, Park et al. [36] analyzed 16 priority
PAHs in the concentration range 0.14 ng/m3 (ACL) to
4.89 (BBF) ng/m3, using GC-MS-based method after
collection on QFFs and extraction in dichloromethane
with ultrasonication. They used external PAH standards
(16 PAHs in a mixture) to quantify the individual ana-
lytes by GC-MS analysis.
Mandalakis et al. [31] monitored 20 PAHs (16 priority
PAHs + Me-phenanthrenes, dimethyl-phenanthrenes,
Me-chrysene, and COR) in samples collected from diverse
locations (an urban center, an inland site and a coastal
area). They analyzed the target PAHs in concentration
range of 0.01 ng/m3 (Me-Chrysene in the inland area) to
7.48 ng/m3 (PHN in the coastal area) by GC-MS analysis
in SIM mode after extracting the filter (GFFs) in hexane
(Soxhlet extraction).
In a similar study, Tsapakis and Stephanou [32]
analyzed 24 particle-bound PAHs from an urban area of
Heraklion, Greece. Extraction of PAHs collected on GFFs
was done with hexane in Soxhlet apparatus, and 24
PAHs (16 priority PAHs + 8 additional ones) were
quantified in a concentration range of 0.02 ng/m3 (FLR)
to 3.27 ng/m3 (BGP) by GC-MS-based analysis in SIM
mode.
Numerous modifications have also been introduced in
the analytical procedures of solvent-based methods.
Chaspoul et al. [38] attempted to analyze 16 priority
PAHs and six nitro-PAHs simultaneously from industrial
workplaces by GC-MS analysis following solvent-
extraction (QFF samples). They used two different
ionization modes in MS [i.e. electron-impact (EI) for
priority PAHs and electron-capture negative ionization
(ECNI) for nitro-PAHs]. These authors were able to
analyze the nitro-PAHs and priority PAHs in the ranges
0–20414 ng/m3 and 65–22180 ng/m3, respectively.
Personal environmental monitors mounted with
Teflon filters were also used for particulate PAHs anal-
ysis from indoor locations divided into smoking and non-
smoking zones in Taiwan [40]. These authors could
quantify airborne PAHs as the geometric mean of all 16
priority PAHs in particle phase (460 ng/m3) through a
GC-FID-based determination.
The analysis of nitro-PAHs and oxy-PAHs was
accomplished by high-resolution GC combined with ion-
trap HRGC-MS2 [34]. This method was effective enough
to monitor nitro-PAHs in the range 2–62 pg/m3, which
was 10–100 times lower than their oxy-PAH counter-
parts.
A two-step supercritical-fluid extraction (SFE) method
was also developed for the analysis of oxy- and nitro-
PAHs present in urban aerosol samples (in Barcelona,
http://www.elsevier.com/locate/trac 1727
1728

Trends
http://www.elsevier.com/locate/trac

Table 4. Applicability of gas chromatography (GC)-based methods in the analysis of polycyclic aromatic hydrocarbons (PAHs) in the airborne particle phase

Target PAHs modea Environmental condition Location Period Detection method Concentrations (ng/m3) Ref.
A
Rural Haven beach, Chesapeake Bay (USA) 1994–1995 GC-MS 7.2 · 10 4 (DBA) to 0.32 (BBF) [29]
Urban Hampton, Chesapeake Bay (USA) 1994–1995 GC-MS 1.98 · 10 3 (DBA) to 0.26 (BBF) [29]
Industrial Elizabeth river, Chesapeake Bay (USA) 1994–1995 GC-MS 7.3 · 10 3 (DBA) to 0.43 (BBF) [29]

B
Urban Rome (Italy) 2002–2003 GC-MS 0.9 (PRL) to 12 (NAP) [47]
Urban Augsburg (Germany) 2002–2005 GC-TOFMS 0.12 (Retene) to 1.06 (sum of benzofluoranthenes) [23]
Urban background Wiesbaden (Germany) 2006–2007 GC-MS 0.03–2.36 (BAP) and 0.03–4.86 (Retene) [59]
Urban bus terminal USA TOF-MS 10 (PAHs) and 0.1 (nitroPAHs) [20]

C
Urban Seoul (Korea) 1998–1999 GC-MS 0.14 (ACN) to 4.89 (BBF) [36]
Urban Campo Grand (Brazil) 1998 GC-MS 8.94 to 62.5 (particle + gas) [17]
Urban Chiang Mai (Thailand) 2004 GC-MS 0.92 to 1.70 (BAP) [39]
Urban Changsha (China) 2008 GC-MS 0.013 (BKF) to 4.85 (PYR) [25]
Sub-urban area Beijing (China) 2005–2006 GC-MS 0.01 (ACN in summer) to 124 (PYR in winter) [55]
Indoor locations Taichung (Taiwan) 2000–2001 GC-FID 460 (Geometric mean of 16 PAHs) [40]
University campus Ghent (Belgium) 2005–2006 GC-MS 0.45 (DBA) to 116 (NAP) [49]

D
Urban Heraklion (Greece) 2000–2002 GC-MS 0.02 (FLT) to 3.27 (BGP) [32]

Trends in Analytical Chemistry, Vol. 30, No. 11, 2011

Urban Basle (Switzerland) 1995 HRGC-MS/MS 3 · 10 4 (Nitro-NAP) to 0.528 (Anthroquinone) [34]
Urban Athens (Greece) 2000 GC-MS 0.05 (COR) to 0.68 (BBF) [31]
Urban Kuala Lumpur (Malaysia) 1998–1999 GC-MS 6.28 (Total PAH) [30]
Industrial France 2003–2004 GC-MS 8.5 (BAA) to 194 (ACN) lg/L [38]
Coastal Chicago (USA) 1994–1995 GC-MS 0.69–45 (mean of 26 PAHs) [58]
Coastal Athens (Greece) 2000 GC-MS 0.01 (DBA, COR, PRL) to 0.37 (CHY) [31]
Background Athens (Greece) 2000 GC-MS 0.03 (ANT) to 0.22 (CHY) [31]
Rural Kuala Lumpur (Malaysia) 1998–1999 GC-MS 0.30 (Total PAH) [30]

E
Urban Barcelona (Spain) 2001 GC-ECD and MS 0.015 to 0.364 [33]
Urban Birmingham (UK) and Damascus (Syria) 1995–1996 GC-MS 0.009 to 0.35 [35]
Rural Maurienne valley (France) 2002–2003 GC-MS 0.13–0.42 (oxPAHs) & 2 · 10 4–7.67 · 10 2
[57]
a
Target mode (A = some of 16 priority PAHs of US EPA, B = some of 16 + some extra, C = 16 PAHs only, D = 16 PAHs + some extra, and E = All the PAHs except 16 PAHs).
 Trends in Analytical Chemistry, Vol. 30, No. 11, 2011
Table 5. Applicability of gas chromatography (GC)-based methods in the analysis of polycyclic aromatic hydrocarbons (PAHs) in the airborne vapor phase

Target PAHs modea Environmental condition Location Period Detection method Concentrations measured (ng/m3) Ref.
A
Rural Haven beach, Chesapeake Bay (USA) 1994–1995 GC-MS 5.8 · 10 4 (BAP) to 6.9 (PHE) [29]
Urban Hampton, Chesapeake Bay (USA) 1994–1995 GC-MS 7.2 · 10 4 (BGP) to 124 (PHE) [29]
Urban Rome (Italy) 2000 GC-MS 231–1743 (NAP), 11–51 (PHN), 2.1–12.1 [15]
(FLT), and 9.2–39
Industrial Elizabeth river, Chesapeake Bay (USA) 1994–1995 GC-MS 1.74 · 10 3 (BKF) to 25 (PHE) [29]

B
Urban Rome (Italy) 2002–2003 GC-MS 0.3 (BAP) to 687 (NAP) [47]
Urban Chicago (USA) 1994–1995 GC-MS 27–430 [58]
Coastal Chicago (USA) 1994–1995 GC-MS 0.08 to 70 [58]

C
Indoor locations Taichung (Taiwan) 2000–2001 GC-FID 1010 (Geometric mean of 16 PAHs) [40]
Urban Campo Grand (Brazil) 1998 GC-MS 8.94 to 62.5 (particle + gas) [17]
Urban Seoul (Korea) 1998–1999 GC-MS 0.80 (BAA) to 15.03 (PHN) [36]
Urban Tainan (Taiwan) 1991–1992 GC-FID 192 (ACL) to 3040 (NAP) [48]
Sub-urban area Beijing (China) 2005–2006 GC-MS 0.79 (BAA in summer) to 2592.5 (NAP in Spring) [55]
University campus Ghent (Belgium) GC-MS 0.45 (DBA) to 116 (NAP) [49]

D
Coastal Athens (Greece) 2000 GC-MS 0.01 (BAP) to 7.48 (PHE) [31]
Background Athens (Greece) 2000 GC-MS 0.01 (Me-CHY) to 1.03 (PHE) [31]
Urban Athens (Greece) 2000 GC-MS 0.01 (BAP) to6.95 (Me-PHE) [31]
Urban Heraklion (Greece) 2000–2002 GC-MS 0.001 (COR) to 19.8 (PHN) [32]
http://www.elsevier.com/locate/trac

E
Urban Birmingham (UK) and Damascus (Syria) 1995–1996 GC-MS 0.01 (9-nitro-ANT) to 0.21 (1-nitro-NAP) [35]
Rural Maurienne valley (France) 2002–2003 GC-MS 0.01–1.43 (oxi-PAHs) & 3 · 10 4–2.2 · 10 2
[57]
a
Target mode (A = some of 16 priority PAHs of US EPA, B = some of 16 + some extra, C = 16 PAHs only, D = 16 PAHs + some extra, and E = All the PAHs except 16 PAHs).

Trends
1729
Trends
Spain) by GC with electron-capture detection (ECD)
coupled to MS [33]. The SFE method, applied to the
analysis of oxy- and nitro-PAHs in urban aerosol sam-
ples, was able to identify 9-fluorenone, 9,10-anthraqui-
none, 2-methyl-9,10-anthraquinone, benzanthrone,
benz[a]anthracene-7,12-23 dione and 1-nitropyrene at a
concentration range of 15–364 pg/m3.
Dimashki et al. [35] used Teflon-coated GFFs to collect
particle-phase PAHs from Birmingham, UK, and
Damascus, Syria. These authors conducted the Soxhlet
extraction in dichloromethane for subsequent GC-MS
analysis and found nitro-PAHs in the range 0.13–
0.42 ng/m3.
4.2. Solvent-based analysis of gas-phase PAHs
Similar to the methods employed in the analysis of par-
ticle-phase PAHs, those developed for gas-phase PAHs
have also relied on extraction of the sorbent material.
Hence, gas-phase PAHs (e.g., PUF) extracted into some
organic solvents can be subject to final determination by
GC-MS.
Gustafson and Dickhut [29] measured gas-phase
PAHs at four sites representative of rural, semi-urban,
urban, and industrialized areas. They collected gas-phase
PAHs on two PUF plugs. Those samples were extracted
in organic solvents (e.g., acetone, petroleum ether, and
dichloromethane) with Soxhlet apparatus and analyzed
for 14 PAHs with deuterated PAH-surrogate standards.
The extracts were then concentrated using rotary
evaporation followed by evaporation under purified
nitrogen. These samples were again extracted with
hexane and cleaned using solid-liquid chromatography
on silica gel for the final determination using GC-MS.
They characterized 14 out of 16 priority PAHs in a
concentration range 0.58 pg/m3 (BAP at rural site) to
124 ng/m3 (PHN) in the urban area.
In another study, Park et al. [36] collected 16 priority
PAHs in gas phase using a PUF sampler in an urban area
of Seoul, Korea. The analysis by a GC-MS method (in SIM
mode) after extraction with dichloromethane (with
ultrasonication) revealed gas-phase PAHs falling in the
range 0.80 ng/m3 (BAA) to 15.03 ng/m3 (PHN).
Mandalakis et al. [31] monitored 20 PAHs (16 priority
PAHs + four others) in samples collected from diverse
locations (urban center, an inland site and the adjacent
coastal area). They could determine target PAHs in the
range 0.01 (PHN) (Me-CHY in the inland area) to
7.48 ng/m3 (PHN in the coastal area) by combining
extraction with hexane (in Soxhlet apparatus) and GC-
MS analysis.
In a similar study, Tsapakis and Stephanou [32] col-
lected samples of 24 PAHs (16 priority PAHs + eight
others) in urban air on PUF, extracted them with hexane
in Soxhlet apparatus, and analyzed them by GC-MS in
SIM mode. They determined them in the range
0.001 ng/m3 (COR) to 19.8 ng/m3 (PHN).
1730 http://www.elsevier.com/locate/trac
Trends in Analytical Chemistry, Vol. 30, No. 11, 2011
Bertoni et al. [15] also analyzed gas-phase PAHs in
concentration range 231–1743 ng/m3 (NAP), 11–
51 ng/m3 (PHN), 2.1–12.1 ng/m3 (FLT), and 9.2–
39 ng/m3 (CHY) in an urban area of Rome, Italy, with a
combination of a passive sampling technique (Carbopack
C as sorbent) and a GC-MS method.
Dimashki et al. (35) analyzed nitro-PAHs in the gas
phase by treatment of a PUF sample in dichloromethane
with Soxhlet apparatus. They determined them in the
range 0.01 ng/m3 (9-nitro-ANT) to 0.21 ng/m3 (1-ni-
tro-NAP) in an urban area of Birmingham, UK, by a GC-
MS analysis in NICI mode.
5. Solvent-free methods for PAH analysis
Solvent-based extraction methods, if employed in the
determination of atmospheric PAHs, generally involve
time-consuming, labor-intensive procedures, regardless
of PAH-phase type. As this step requires the use of toxic
organic solvents, it may cause added difficulties with
sample handling during the extraction stage [62]. To
overcome such limitations, solvent-free methods (e.g.,
TD) have commonly been employed. Their applicability
has been demonstrated in the analysis of VOCs and re-
duced sulfur compounds (RSCs) [63]. As for solvent-
based methods, the applicability of this technique may
also be considered separately for gas-phase and particle-
phase PAHs.
5.1. Solvent-free analysis of particle-phase PAHs
Two types of filters (i.e. QFF and Teflon) have commonly
been utilized for the collection of the particle-bound
PAHs that are subsequently retrieved directly without
using solvent.
As an example of the use of a Teflon filter for solvent-
free collection device, Bezebeh et al. [64] analyzed
aerosol-particulate matter for some of the 16 priority
PAHs and nitro-derivatives using laser-desorption-
ionization time-of-flight MS (LDI-TOF-MS). They col-
lected aerosol particulate matter (PM2.5) on Teflon filters
from inside a bus terminal. The analytes were introduced
into the LDI-TOF-MS by mounting a 3-mm diameter
piece of the filter on a probe. The probe was rotated to
allow the analysis of samples at multiple positions on the
aerosol filter. The collection of a small sample volume
(0.32–0.98 m3) and desorption and ionization with
pulsed UV radiation at 266 nm resulted in detection of
PAHs in the positive-ion spectra, while nitro-PAHs were
detected in the negative-ion spectra. By applying this
technique, they were able to quantify several priority
PAHs (PYR, CHY, TPN, PRL, and DBA) and three nitro-
PAHs at nearly 10 ng/m3 and 0.1 ng/m3, respectively.
QFFs have been employed frequently to collect par-
ticle-bound PAHs for subsequent solvent-free analysis.
PM2.5 samples collected on QFFs from an urban area of
Trends in Analytical Chemistry, Vol. 30, No. 11, 2011
Germany (Augsburg city) were analyzed for many of
the 16 priority PAHs and oxy-PAHs by direct thermal
desorption (DTD)-GC-TOF-MS [23]. For this purpose,
the filter samples were cut into pieces (2.1 · 27 mm).
Two filter sub-samples, each of which represented
1.1 m3 of sampled air, were placed into a GC-liner.
Isotope-labeled reference compounds were added (for
quantification), and liners were put into the cold
injector. Desorption of target PAHs was accomplished
at 320C (for 15 min). Based on this method, these
authors quantified the target PAHs in the sub-ng/m3
range.
In a similar study, a two-step TD injection system
integrated into a GC-MS was applied as a solvent-free
method for the determination of PAHs across a wide
volatility range (PHN to BGP) from ambient air particles
[59]. These authors collected PM10 samples on QFFs at
high volume (720 m3 at 1250 L/min) and low volume
(6.9 m3 at 38.3 L/min) from an urban location in
Wiesbaden, Germany. These QFF samples were then cut
into pieces, and one piece (corresponding to a high
sample volume of 180 m3) was placed in a glass liner.
After adding an internal standard, the filter piece was
thermally desorbed in a TD unit interfaced with a GC-MS
system. These authors determined PAHs in the concen-
tration ranges of 0.03–2.36 ng/m3 (BAP) and 0.03–
4.86 ng/m3 (Retene).
Likewise, Gil-Molto et al. [61] relied on a TD method to
analyze 12 (out of 16) priority PAHs on QFFs (in PM10
and PM2.5 samples), collected by low-volume samplers
from the city of Elche, Spain. These filter samples were
rolled and placed in TD tubes to desorb the target PAHs
at 300C (for 10 min) in an automatic TD system
interfaced with GC-MS. These authors successfully
determined the PAHs in both PM2.5 and PM10 samples in
the concentration ranges 0.020 ng/m3 (DBA) to 0.41
ng/m3 (CHY) and 0.002 ng/m3 (ANT) to 0.94 ng/m3
(BBF) , respectively.
The reliability of the TD-GC-MS technique was eval-
uated for 4–6 ring PAHs collected on PM10 quartz filters
[60]. Their TD device was equipped with a desorption
oven connected to a Peltier-cooled, sorbent-packed, cold-
trapping system. Through application of a two-stage
focusing technique, they were able to achieve complete
recovery of samples of a NIST urban dust [Standard
Reference Material (NIST SRM 1649a)] at the 95%
confidence level based on the certified values supplied
(except for BAA).
For the speciation of organic components, including
PAHs in ambient air, PM2.5 samples were analyzed by a
TD-GC-MS method [65]. The authors collected PM2.5
samples on QFFs and cut two separate punches (8.1 mm
in diameter, 0.518 cm2 in area) to load into the TD glass
tube. The samples were then analyzed via thermal
desorption combined with GC-MS. Based on this method,
these authors quantified the PAHs in PM2.5 samples in
Trends
the range 0.02 ng/m3 (FLR) to 1.64 ng/m3 (ICP) from
an urban area (Golden BC, Canada).
Falkovich and Rudich [41] reported a direct sample-
introduction (DSI) method for filter or liquid samples into
a GC injector for thermal desorption and subsequent GC-
MS analysis to identify a number of semi-VOCs, includ-
ing 16 priority PAHs. Their sampling was conducted on
aluminum foil (preferred over QFFs by these authors) to
be desorbed in micro-vials inside the GC injector (tem-
perature 400–450C). For quantification of target PAHs,
2 lL of the PAH standard solutions (EPA 610 mixture)
contained in micro-vials were thermally desorbed and
similarly analyzed on aluminum foil. In this way, they
were able to quantify PAHs in the concentration range
0.005 ng/m3 (PYR) to 0.34 ng/m3 (PHN) in PM10
samples collected from an urban area in Tel Aviv, Israel.
In an effort to improve a solvent-free quantification
method for particle-phase PAHs, Waterman et al. [66]
conducted a TD-GC-MS analysis of NIST SRM1649a
(urban dust). They prepared samples with a known
quantity of urban dust in a glass-lined, stainless-steel GC
liner (cleaned at 600C for 5 min) for thermal desorp-
tion. The samples vaporized from the liner were then
focused on a cold trap ( 196C with liquid nitrogen)
and flash heated from 196 to 300C for 20 s for sub-
sequent GC-MS analysis. They were able to confirm that
the values that they obtained agreed with the certified
values at the 95% confidence interval.
As a solvent-free technique, direct HS-SPME (with a
PDMS fiber) was also developed for the determination of
particulate PAHs collected on Teflon-coated GFF [67].
HS-SPME extractions were performed in 0.6 mL of 20%
sodium sulfate in deionized water. The extraction was
conducted in two steps: (1) equilibration of the filter
(with PAH sample) and the liquid phase (sodium sulfate
in water) and (2) SPME extraction at 50C for 120 min.
These authors were able to quantify the 16 priority
PAHs in the concentration range of 0.20 (ANT) to
1.8 ng/m3 (PYR) in PM10 samples collected from an
urban area in Brazil (Sao Paulo city) by GC-MS analysis.
5.2. Solvent-free analysis of gas-phase PAHs
For solvent-free analysis of gas-phase PAHs, many types
of sorbent materials have been tested and evaluated,
either individually or in combinations. Baltussen et al.
[68] proposed an analytical method for gas-phase PAH
analysis based on enrichment of the solutes on a packed
bed of 100% PDMS particles, followed by TD-GC-MS. The
PDMS material exhibited excellent thermal stability even
after 200 consecutive runs. Using the above method,
they were able to determine 16 priority PAHs in the
concentration range of 0.50 ng/m3 (BAP) to 129 ng/m3
(ACL).
Recently, a novel analytical method for atmospheric
PAHs was developed based on laser-induced fluorescence
(LIF) of samples on quartz multi-channel PDMS traps
http://www.elsevier.com/locate/trac 1731
Trends
with the aid of diffusion-based sampling. The method
allowed the rapid (<5 min), cost-effective analysis of
samples [62]. However, small amounts of naphthalene-
photodegradation products (phenol, benzyl alcohol, and
phthalic anhydride) were also identified by TD-GC-MS
after >15 min irradiation.
MAD coupled to headspace solid-phase micro-
extraction (HS-SPME) has been studied for in-situ, one-
step, sample-preparation step for PAHs collected on
XAD-2 adsorbent for subsequent GC-MS analysis [18].
The PAHs on XAD-2 were desorbed into the extraction
solution (ethylene glycol), evaporated into the headspace
by microwave irradiation, and absorbed directly on a
SPME fiber in the headspace. PAHs were then analyzed
by desorbing the fiber in the injection port. This method
was applied to collect smoke samples from the indoor
burning of joss sticks [18]. The quantities of PAH
determined via this approach were in the range 0.795–
2.53 ng.
Wauters et al. [49] used a mixed sorbent bed (i.e.
PDMS foam, PDMS particles, and a Tenax TA bed) for
the collection of 16 priority PAHs in the gas phase.
Quantitative analysis of those samples was then made by
TD-GC-MS. They also compared their method to the
classical method that typically relies on the high-volume
sampling (on GFF followed by PUF) for 24 h. These au-
thors could monitor the 16 priority PAHs in the con-
centration range 0.45 ng/m3 (DBA) to 116 ng/m3
(NAP) from the sample collected inside the university
campus at Ghent, Belgium.
A sensitive, solvent-free method was also developed to
quantify personal exposure to gaseous PAHs (NAP, ACN,
ACL FLR, PHN, ANT, FLT, and PYR) via passive sam-
pling with a lab-made sampler [50]. They were then able
to analyze gas-phase PAHs in the concentration range
1.9–2600 ng/m3 by a TD-GC-MS-based analysis.
6. Evaluation of basic quality assurance in PAH
analysis
Considering that airborne PAHs are present at very low
concentrations (at or below the ng/m3 range), the most
important parameter for the basic quality assurance
(QA) is LOD, which is often expressed in terms of abso-
lute mass (e.g., ng) and/or mass per volume (e.g.,
ng/m3). Although most of the studies dealing with PAH
analysis relied on GC-MS, the magnitude of LOD values
varied noticeably, depending upon the analytical con-
ditions and instrumental set-ups. Moreover, it also varied
between the PAH phases (e.g., gas or particle). It is also
an important consideration that, because GC-MS is a
highly specific technique, whereby an analyte is only
detected if it elutes from the GC at the right time, and, in
addition, exhibits the molecular ion fragment being
monitored in MS, LODs based on the standard formula of
1732 http://www.elsevier.com/locate/trac
Trends in Analytical Chemistry, Vol. 30, No. 11, 2011
three times the standard deviation of the blank responses
are sometimes very difficult to calculate. On occasions
where the blank response is too low, other methods to
calculate the LODs based on fit-for-purpose uncertainty
assessment, and sample adulteration with impurities
have been used [69]. If the LOD values are to be ex-
pressed in concentration terms to yield a method LOD,
the volume of total samples is also a critical variable to
assess. In addition, the LOD will be affected by many
other method parameters (e.g., the reduced volume of
the final extract and the volume injected into the GC-MS)
[70].
To provide an overview of the normal detection range
of PAHs across different analytical approaches, the LOD
values of different studies are compared in Table 6
(particle) and Table 7 (vapor phase). As shown in
Table 6, the LOD values for particle-phase PAHs varied
over the ng to sub-ng range. Although these LOD values
are also variable across individual PAHs, the data for
each PAH are rather scarce. In addition, most studies
reported a range of LOD values for their analysis instead
of each individual one.
Simcik et al. [58] calculated LOD values for 10 target
PAHs (seven of 16 priority + three extra) in the particle-
phase and reported their LOD range as 0.002–59 ng
based on GC-MS analysis (after solvent treatment).
Comparatively high LOD values were often reported for
some lighter PAHs with high volatility (e.g., NAP).
Possanzini et al. [47] reported an LOD value of 5.9 ng for
NAP in the particulate phase, which was an order of
magnitude greater than their ANT value (0.3 ng); these
authors also relied on a solvent-based GC-MS analysis.
Although a wide variation in LOD values was reported
for many individual PAHs, some authors also found a
relatively narrow range of LODs across different indi-
vidual PAHs (Table 6). Van Drooge et al. [59] could
achieve a limit of quantification (LOQ) value near
0.08 ng for most of the 16 priority PAHs in particle
phase, based on a two-step TD and GC-MS method. In a
similar study, Gil-Molto et al. [61] also found their LOD
values in the sub-ng range (0.01–0.5 ng) for 16 priority
PAHs in particle phase by TD-GC-MS analysis.
According to our assessment of a large number of LOD
values, the values for vapor-phase PAHs generally fell
near, or below, the ng range. Lee et al. [48] reported LOD
values for 16 priority PAHs in the range 0.07–0.15 ng
with a solvent-based GC-MS analysis. In another study,
Wei et al. [18] reported LOD values in the range 0.02–
1 ng using an HS-SPME-GC-MS method. Possanzini et al.
[47] also found relatively high LOD values in vapor
phase [i.e. 1 ng (CHY) to 15.3 ng (NAP) by a GC-MS
method (solvent-based)]. Likewise, Simcik et al. [58]
found a comparatively wide range of LODs for vapor-
phase PAHs (i.e. 0.002–1406 ng) in a solvent-based GC-
MS analysis. Apart from these PAHs, Albinet et al. [57]
achieved LOD values for PAH derivatives in sub-pg
 Trends in Analytical Chemistry, Vol. 30, No. 11, 2011
Table 6. Compilation of basic quality assurance (QA) data for the analysis of airborne polycyclic aromatic hydrocarbons (PAHs) in the vapor phase

Target PAHs Extraction (Solvent- Pretreatment and Detection LOD Precision Recovery Ref.
modea based or Solvent free) sample loading method
Flow rate Sampling Absolute Concentration (% RSD) (%)
(L/min1) volume (L) mass (ng) (ng/m3)
A
4
Soxhlet extraction with DI GC-MS 510–750 172,000– 1.2–7.9 · 10 79.3 to 103 [29]
acetone, ether, and 665,000
dichloromethane
No Thermal desorption GC-MS 0.3 [62]
No Laser induced GC-MS 1000 [62]
fluorescence (LIF)
Ethylene glycol Microwave assisted GC-MS 1.4 420 0.02 to 1 4 to 14 80–108 [18]
desorption & HS-SPME
(PDMS-DVB fiber)
No Thermal desorption GC-MS 12 to 25 [50]
Toluene DI GC-MS 1.0 94–100, 85–87, [15]
61–77, and 53–59

B
Soxhlet extraction with DI GC-MS 6 36 1 (CHY) to 4 to 9 90–97 [47]
dichloromethane– 15.3 (NAP)
methanol
Soxhlet extracted with DI GC-MS 0.002–1460 78 to 101 [58]
dichloromethane

C
Hexane:methylene DI GC-FID 4 32 0.17 to 0.49 lg/mL below 10 85–115 [40]
chloride with sonication (except PHE, 20)
Dichloromethane- DI GC-MS 34 146880 [17]
methanol with
ultrasonication
Hexane DI GC-MS 2.5 60 1–10 ppb below 5 95.1 [16]
Dichloromethane with DI GC-MS 83 [36]
ultrasonication
http://www.elsevier.com/locate/trac

Soxhlet with hexane and DI GC-MS 1.3 624 0.3 (BAA) 63.2 (Nap) to [55]
cylcohexane to 170.5 (NAP) 108 (BaP)
Soxhlet extraction with GC-FID 0.071 to 0.153 0.6 to 6.7 [48]
mixture of n-hexane,
acetone, and
dichloromethane in
2:1:1 mixture)
PDMS traps TD GC-MS 0.1 [68]
Sorption tubes (PDMS TD GC-MS 0.1 144 0.007 (ANT) 3.7 (PYR) to 80. 5 (NAP) to [49]
foam + PDMs to 0.026 (NAP) 12.9 (NAP) 118 (BGP)
particle + Tenax bed)

Trends
1733

(continued on next page)

Trends Trends in Analytical Chemistry, Vol. 30, No. 11, 2011

(oxy-PAHs) to sub-ng range (nitro-PAHs) in both parti-

[31]

[32]

[35]

[57]
Ref.
cle and gas phase through application of a solvent-based

Target mode (A = some of 16 priority PAHs designated by US EPA, B = some of 16 + some extra, C = 16 PAHs only, D = 16 PAHs + some extra, and E = All the PAHs except 16 PAHs).
GC-MS method.

14–84 (NitroPAHs)
Although the absolute LOD reflects the most impor-

& 5-83 (oxPAHs)

Recovery

tant feature of any analytical method, it can be manip-

(%)

ulated to a certain degree by altering the initial volume

of air sampled. Hence, for comparison of LODs in a
practical sense, they should be compared in terms of
both absolute mass and mass concentration (e.g.,
Precision
(% RSD)

ng/m3). If the available research is surveyed from this

aspect, most studies were able to quantify PAHs at
ng/m3 or even pg/m3 levels (Tables 6 and 7).
As another important QA parameter, reproducibility
Concentration

expressed in terms of relative standard deviation (RSD%)

(ng/m3)

can be examined across different studies. If the data on

0.001

PAHs are compared from this aspect, they vary greatly

across individual compounds – as low as 2% in case of
FLT and as high as 28% for ANT in a TD-GC-MS analysis
0.03-0.07 (nitroPAHs)

of particle-phase PAHs in NIST SRM1649a (urban dust)

mass (ng)

[61]. Likewise, it was below 10% for all the 16 priority

Absolute

(ox PAHs) and

0.01-2.60 pg

PAHs, except for PHN (20%) in both particle and gas

LOD

phases, when analyzed by a solvent-based GC-MS

method for indoor-air samples [40]. Interestingly, in
vapor phase, the lowest values were 3% (PHN), while the
highest was 14% (PYR) [18]; these authors used a
volume (L)
Sampling

1000,000
450,000–

900,000–

microwave-assisted HS-SPME-GC-MS method.

900,000
900,000

Wauters et al. [49] assessed the reproducibility of their

TD-GC-MS analysis for 16 priority PAHs and found them
in the range 3.7% (PYR) to 12.9% (NAP), with an
Flow rate
(L/min1)

average of 7.4%. Hence, the reproducibility of PAH

1250
600

analysis is generally seen to be maintained near or below

the 10% level (RSD). However, on certain occasions, this
Detection

TD-GC-MS method suffered from relatively poor repro-

method

GC-MS

GC-MS

GC-MS

GC-MS

ducibility, especially for lighter PAHs, regardless of phase

type.
Although still fairly rare, some studies aimed to com-
pare the relative performance between emerging TD-
Cool splitless injection
Pretreatment and
sample loading

based methods and conventional solvent-based methods

as a means to warrant reliability in PAH determination.
Wauters et al. [49] compared their TD-GC-MS method to
(40–320C)

the classical method for quantitative analysis of the 16

priority PAHs. They found that concentrations measured
DI

DI

DI

using the new method were significantly higher than

those obtained using the classical method (i.e. a factor
dichloromethane/Liquid
based or Solvent free)

Soxhlet extraction with

Soxhlet extraction with

Soxhlet extraction with

1.2–3 for the high molecular weight PAHs). Moreover,

Extraction (Solvent-

purification with SPE

those differences were extended further for lighter PAHs

Pressurized liquid
dichloromethane

chromatography

up to 35 times for NAP and 23 times for ACL.

extraction with

In another study, Van Drooge et al. [59] assessed the

cartridges

detectability of both conventional solvent extraction and

hexane

hexane

a new TD-based method interfaced with a GC-MS system.

Table 6. (continued)

The TD method was reported to have lower uncertainties

than the conventional solvent-based extraction method,
Target PAHs

if the analysis of more reactive PAHs is made at similar

modea

concentrations (e.g., BAP); the use of TD methodology

instead of the solvent-based method enabled the reduc-
D

tion of the analytical expanded uncertainty from 19% to

1734 http://www.elsevier.com/locate/trac
 Trends in Analytical Chemistry, Vol. 30, No. 11, 2011
Table 7. Compilation of basic quality assurance (QA) data for the analysis of airborne polycyclic aromatic hydrocarbons (PAHs) in the particle phase

Target PAHs Solvent-based/Solvent Pretreatment and Detection LOD Precision Recovery Source
modea free Extraction (Solvent- sample loading method
based or Solvent free) Flow rate Sampling Absolute Concentration (% RSD) (%)
(L/min) volume (L) mass (ng) (ng/m3)
A
Solvent free TD GC-MS 2–6 mg/kg [66]
4
Soxhlet extraction with DI GC-MS 510–750 172,000 to 1.2–7.9 · 10 79.3 to 103 [29]
acetone, ether, and 665,000
dichloromethane

B
Soxhlet extraction with DI GC-MS 6 36 0.3 (ANT) to 4 to 9 [47]
dichloromethane– 5.9 (NAP)
methanol
Soxhlet-extraction with Flash chromatography GC-MS 1250 720,000 8 81 [37]
dichloromethane/n- with silica gel + toluene/
hexane hexane as solvent
No 2 step thermal GC-MS 38.3 and 1250 6900 and 0.08–4.0 8 to 16 [59]
desorption 720,000 (analytical
uncertainty)
No TD GC-MS 0.01–0.05 3.84 to 14.76 more than 95 [61]
No TD GC-MS 38.3 0.01–0.05 2 to 28 (Anth) more than 95 [61]
Soxhlet extracted with DI GC-MS 0.002–59 78 to 101 [61]
dichloromethane
No Laser desorption TOF-MS 2.7 320–980 0.01 [61]
ionization (LDI)

C
Dichloromethane with DI GC-MS 88 [36]
ultrasonication
Hexane:methylene DI GC-FID 4 0.17 to 0.49 lg/mL below 10 85–115 [40]
http://www.elsevier.com/locate/trac

chloride with sonication (except Phe, 20)

Dichloromethane- DI GC-MS 34 146880 [17]
methanol with
ultrasonication
Soxhlet extraction with GC-MS 100 86400 0.001 (DBA) to 84.5 (NAP) to [55]
hexane and cyclohexane 0.74 (NAP) 133.1 (BAP)
Acetonitrile with DI GC-FID 1.6 (CHY) to [39]
utltrasonication 6.2 (ICP)
Acetonitrile with DI GC-MS 0.12 (FLR) to 44 (NAP) to [39]
utltrasonication 0.45 (ANT) 121 (DBA)
Ultrasonic agitation in DI GC-MS 100 48000 0.003 (ACL) to 73–88 [25]
dichloromethane 0.508 (PHE)

Trends
1735

(continued on next page)

1736

Trends
Table 7. (continued)

Target PAHs Solvent-based/Solvent Pretreatment and Detection LOD Precision Recovery Source
modea free Extraction (Solvent- sample loading method
http://www.elsevier.com/locate/trac

Flow rate Sampling Absolute Concentration (% RSD) (%)

based or Solvent free)
(L/min) volume (L) mass (ng) (ng/m3)
D
Soxhlet extraction with DI GC-MS 600 900,000 0.001 [32]
hexane
Soxhlet extraction with DI GC-MS 1 240–480 4.2 to 16 [38]
dichloromethane (Coefficient
of variation)
Microwave assisted DI HRGC-MS/MS 666 960,000 80–90 [40]
methanol-acetone
Soxhlet extracted with DI GC-MS 0.002–59 78 to 101 [58]
dichloromethane
Soxhlet extraction with DI GC-MS 450,000– [31]
hexane 900,000
Dichloromethane with DI GC-MS 1120 1612,000– 50.77 to 99.70 [30]
ultrasonication 3224,000

E
Supercritical fluid GC-ECD 1000 1,440,000 [33]
extraction and MS
Soxhlet extraction with Direct injection (DI) GC-MS 900,000– [35]
dichloromethane 1000,000
Pressurized liquid Cool splitless injection GC-MS 1250 0.01–2.60 pg 14–84 (NitroPAHs) [57]
extraction with (40–320C) (ox PAHs) and & 5-83 (oxPAHs)
dichloromethane/Liquid 0.03–0.07

Trends in Analytical Chemistry, Vol. 30, No. 11, 2011

chromatography (nitroPAHs)
purification with SPE
cartridges
a
Target mode (A = some of 16 priority PAHs designated by US EPA, B = some of 16 + some extra, C = 16 PAHs only, D = 16 PAHs + some extra, and E = All the PAHs except 16 PAHs).
Trends in Analytical Chemistry, Vol. 30, No. 11, 2011
8% [59]. Based on the initial results of TD methods, it
appears that TD can be a superior choice for certain
PAHs; however, acquiring good recovery of all the major
individual PAHs is still a challenge, as it is sensitive to
the physicochemical properties of individual PAHs.
In fact, quality control (QC) parameters have been
considered in the European standard method for ben-
zo[a]pyrene analysis [56]. Indeed, to highlight the diffi-
culty of these measurements, it is worth noting that the
target relative expanded uncertainty for these measure-
ments is 50% – of which only about one tenth is
attributable to the air sampling. Furthermore, this doc-
ument states that, even if laboratories are highly profi-
cient at an individual analysis technique, it is still
possible to find significant ‘‘between-laboratory’’ vari-
ability between expert laboratories using the same or
different methods. This may contribute to the overall
uncertainty of the method. Full uncertainty budgets for
the GC-MS analysis of PAHs and similar organic com-
pounds are available [71]. The key QC parameters
highlighted in the European reference method EN 15549
are:
(1) Reagent-blank checks (of solutions and collec-
tion media) to ensure blank levels remain accept-
ably low.
(2) Calibration-drift check. Following calibration
with a standard of known composition to establish
the sensitivity of the analyzer, it is recommended
that a representative calibration standard is re-
measured every 10 samples to check for drift in
the analyzer response. If necessary, a drift correc-
tion may be performed.
(3) Measurement of QC solutions on a regular basis
and plotting the results on a QC chart to ensure that
the methods remains under control, and to take
corrective action, if necessary.
(4) Assessing the performance of the method by
regularly measuring reference materials (e.g., NIST
SRM 1649b or the newly released European Union
reference material ERM-CZ100)
(5) Correction of the reduced extract volume by
means of a known quantity of added internal stan-
dard (usually a deuterated PAH). This accounts for
the volume of the extract following rotary evapora-
tion or similar techniques being unlikely to be ex-
actly the same every time. Ratioing of the
analytical response to the response obtained from
the internal standard corrects for this.
(6) Use of surrogate standards to correct for losses
during extraction and sample preparation (as de-
scribed earlier).
(7) Regular participation in interlaboratory com-
parison to assess competency and external quality
assessment by relevant accreditation bodies (e.g.,
UKAS in the UK).
Trends
7. Summary and concluding remarks
Considering the potential health impact of airborne
PAHs, an accurate assessment of their concentration
levels in target environmental settings is crucial to
establish their proper management, to assess abatement
strategies and to develop future policy options. As the
physicochemical properties of PAHs vary between lighter
and heavier compounds, they are fractionated differen-
tially between gas phase (lighter with high volatility)
and particle phase (relatively heavy PAHs). As such, a
suite of PAHs can exhibit noticeable differences in con-
centration levels in line with their molecular weights.
For this reason, it has often been a challenge to develop
precise techniques to measure their distribution in a
particular environment.
In line with technological advancement in analytical
methodologies, we can quantify airborne PAHs, along
with other harmful air contaminants, at trace level
quantities (at or below ng/m3). As is the case for the
most common organic contaminants (e.g., VOCs), GC-
based methods are by far the most reliable for the
analysis of airborne PAHs with a selected preconcen-
tration (or treatment) method. However, due to their
semi-volatile nature, high molecular weight, and differ-
ent reactivities across gas or particle phase, the PAHs
cannot be analyzed without establishing proper analyt-
ical protocols for their final determination with robust
QA procedures. This involves the collection of gas-phase
PAHs on sorbents (e.g., PUF) or particle-phase PAHs on
some filter materials (GFF or QFF) preferably in different
size fractions (i.e. PM10 or PM2.5) relevant to exposure
mechanisms (e.g., inhalation). To this end, sufficiently
large volumes of air are required to pass through high-
volume samplers to collect enough material to analyze
with fit-for-purpose uncertainty.
After collecting PAHs on sorbent or filter, extraction is
normally accomplished with a number of organic sol-
vents with Soxhlet or ultrasonication treatments. These
conventional solvent-based methods have been success-
fully used to retrieve a wide range of PAHs [16 priority
PAHs + other PAHs + their derivatives (e.g., nitro and
oxy)] from various environmental settings. However, the
extraction efficiencies are often highly uncertain, and
those values presented are rarely comparable across
different studies and different methods. The most
noticeable variation in recovery rates of PAHs is ob-
served most commonly for lighter ones or PAHs with
high volatility (e.g., NAP). Figs. 1 and 2 present common
combinations of sampling and analysis strategies em-
ployed by previous researchers for PAHs in particle
phase and gas phase, respectively.
Due to the complexities involved in the multi-step
procedure of sample preparation (for PAHs), the use of
toxic organic solvents, and eventually the probable
http://www.elsevier.com/locate/trac 1737
Trends
chances of analyte loss, there have been a number of
efforts to develop solvent-free analysis of PAHs. These
solvent-free methods rely on the collection of the air-
borne PAHs on sorbent or filter materials for subsequent
release by the thermal desorption. Although these
methods are still in development, initial results have
shown a number of advantages (e.g., improved recovery
and reduced analytical uncertainty) over conventional
methods, especially for lighter PAHs (e.g., NAP and
ACL). Moreover, these methods do not require the high
sample volume (or sampling time), that is a prerequisite
for conventional methods.
For the final determination of these airborne PAHs in
gas and particle phases, GC-MS has been most com-
monly selected. When we compared the LODs of PAH
analysis across different studies, the GC-MS method was
reliable to determine the concentrations of PAHs and
their derivatives at the sub-ng level (in absolute terms),
irrespective of the sample-preparation method. Although
the LODs of all individual PAHs varied to a certain de-
gree across different studies, most of these studies were
able to quantify airborne PAHs at sufficiently low con-
centrations (e.g., pg/m3 level in remote rural, coastal, or
pristine areas). However, they have also shown many
discrepancies in terms of recovery or reproducibility,
depending mainly on the nature of the individual PAHs
in association with possible bias in their collection and/
or sample-preparation steps.
In recent years, many efforts have been made to test
the compatibility between conventional solvent-based
methods and emerging TD-based methods. These com-
parisons have often shown the superiority of TD-based
methods [e.g., reduced analytical uncertainty, improved
recovery, requirement for smaller sample volumes (or
shorter times), and no use of toxic organic solvents].
However, these methods are confined to a relatively
narrow window of target PAHs relative to the conven-
tional ones. Hence, future research should warrant
evaluation of suitable sorbent materials with such con-
siderations as: expansion of the method to cover more
target PAHs and enhanced breakthrough capacity.
Moreover, one also needs to establish some effective
measures to check the accuracy of results between
conventional solvent-based methods and emerging sol-
vent-free methods. This would probably help us under-
stand the atmospheric behavior of these toxic substances
in a more meaningful way and offer a stronger basis to
establish strategies for their management.
Acknowledgements
This work was supported by the National Research
Foundation of Korea (NRF) Grant funded by the Ministry
of Education, Science and Technology (MEST) (No.
2010-0007876).
1738 http://www.elsevier.com/locate/trac
Trends in Analytical Chemistry, Vol. 30, No. 11, 2011
References
[1] R. Barro, J. Regueiro, M. Llompart, C. Garcia-Jares, J. Chromatogr.,
A 1216 (2009) 540.
[2] F.J. Santos, M.T. Galceran, Trends Anal. Chem. 21 (2002) 672.
[3] D.L. Poster, M.M. Schantz, L.C. Sander, S.A. Wise, Anal. Bioanal.
Chem. 386 (2006) 859.
[4] L.A. Gundel, V.C. Lee, R.R.M. Kariawasam, R.K. Stevens, J.M.
Daisey, Atmos. Environ. 29 (1995) 1719.
[5] M. Kallio, T. Hyotylainen, M. Lehtonen, M. Jussila, K. Hartonen,
M. Shimmo, M.L. Riekkola, J. Chromatogr., A 1019 (2003) 251.
[6] K. Dzepina, J. Arey, L.C. Marr, D.R. Worsnop, D. Salcedo, Q.
Zhang, T.B. Onasch, L.T. Molina, M.J. Molina, J.L. Jimenez, Int. J.
Mass Spectrom. 263 (2007) 152.
[7] G.L. Stroher, N.R. Poppi, J.L. Raposo, J.B.G. de Souza, Microchem.
J. 86 (2007) 112.
[8] Y. Tasdemir, F. Esen, Atmos. Res. 84 (2007) 1.
[9] M.C.S. Machado, J. Loyola, S.L. Quiterio, G.O. da Rocha, J.B. de
Andrade, G. Arbilla, J. Brazil Chem. Soc. 20 (2009) 1565.
[10] S.O. Baek, M.E. Goldstone, P.W.W. Kirk, J.N. Lester, R. Perry,
Chemosphere 22 (1991) 503.
[11] N. Poor, R. Tremblay, H. Kay, V. Bhethanabotla, E. Swartz, M.
Luther, S. Campbell, Atmos. Environ. 38 (2004) 6005.
[12] N. Vardar, Y. Tasdemir, M. Odabasi, K.E. Noll, Sci. Total Environ.
327 (2004) 163.
[13] K. Ravindra, L. Bencs, E. Wauters, J. de Hoog, F. Deutsch, E.
Roekens, N. Bleux, P. Bergmans, R. Van Grieken, Atmos. Environ.
40 (2006) 771.
[14] L. Pozzoli, S. Gilardoni, M.G. Perrone, G. de Gennaro, M. de
Rienzo, D. Vione, Ann. Chim. 94 (2004) 17.
[15] G. Bertoni, R. Tappa, A. Cecinato, Chromatographia 53 (2001)
S312–S316.
[16] K.L. Liu, W.J. Han, W.P. Pan, J.T. Riley, J. Hazard Mater. 84
(2001) 175.
[17] N. Re-Poppi, M. Santiago-Silva, Atmos. Environ. 39 (2005) 2839.
[18] M.-C. Wei, W.-T. Chang, J.-F. Jen, Anal. Bioanal. Chem. 387
(2007) 999.
[19] C. Conolly, Annual Report for 2009 on the UK PAH Monitoring
and Analysis Network, AEAT/ENV/R/3059, AEA, Harwell, UK,
2010.
[20] D.Z. Bezabeh, A.D. Jones, L.L. Ashbaugh, P.B. Kelly, Aerosol Sci.
Tech. 30 (1999) 288.
[21] M. Shimmo, H. Adler, T. Hyotylainen, K. Hartonen, M. Kulmala,
M.L. Riekkola, Atmos. Environ. 36 (2002) 2985.
[22] E. Schnelle-Kreis, M. Sklorz, A. Peters, J. Cyrys, R. Zimmermann,
Atmos. Environ. 39 (2005) 7702.
[23] E. Schnelle-Kreis, W. Welthagen, M. Sklorz, R. Zimmermann, J.
Sep. Sci. 28 (2005) 1648.
[24] Y. Kojima, K. Inazu, Y. Hisamatsu, H. Okochi, T. Baba, T. Nagoya,
Asian J. Atmos. Environ. 4 (2010) 1.
[25] F. Yang, Y.B. Zhai, L. Chen, C.T. Li, G.M. Zeng, Y.D. He, Z.M. Fu,
W.F. Peng, Atmos. Res. 96 (2010) 122.
[26] J.M. Delgado-Saborit, N. Aquilina, S. Baker, S. Harrad, C.
Meddings, R.M. Harrison, Anal. Methods 2 (2010) 231.
[27] H.T. Nguyen, K.-H. Kim, C.-J. Ma, J.-M. Oh, Sci. World J. 10
(2010) 20.
[28] A. Anthwal, K. Jung, H.-J. Kim, I.-S. Bae, K.-H. Kim, FreseneniusÔ
Environ. Bull. 19 (2010) 1356.
[29] K.E. Gustafson, R.M. Dickhut, Environ. Sci. Technol. 31 (1997)
140.
[30] N. Omar, M.R. Abas Bin, K.A. Ketuly, N.M. Tahir, Atmos.
Environ. 36 (2002) 247.
[31] M. Mandalakis, M. Tsapakis, A. Tsoga, E.G. Stephanou, Atmos.
Environ. 36 (2002) 4023.
[32] M. Tsapakis, E.G. Stephanou, Environ. Pollut. 133 (2005) 147.
Trends in Analytical Chemistry, Vol. 30, No. 11, 2011
[33] P. Castells, F.J. Santos, M.T. Gaiceran, J. Chromatogr., A 1010
(2003) 141.
[34] M. Niederer, Environ. Sci. Pollut. Res. 5 (1998) 209.
[35] M. Dimashki, S. Harrad, R.M. Harrison, Atmos. Environ. 2000
(2000) 2459.
[36] S.S. Park, Y.J. Kim, C.H. Kang, Atmos. Environ. 36 (2002) 2917.
[37] A. Lottmann, E. Cade, M.L. Geagea, O. Delhomme, C. Grand, C.
Veilleraud, A.L. Rizet, P. Mirabel, M. Millet, Anal. Bioanal. Chem.
387 (2007) 1855.
[38] F. Chaspoul, G. Barban, P. Gallice, Polycycl. Aromat. Compd. 25
(2005) 157.
[39] S. Chantara, W. Sangchan, ScienceAsia 35 (2009) 42.
[40] S.C.C. Lung, M.J. Wu, C.C. Lin, J. Expo. Anal. Environ. Epidemiol.
14 (2004) 529.
[41] A.H. Falkovich, Y. Rudich, Environ. Sci. Technol. 35 (2001) 2326.
[42] M. Tsapakis, E.G. Stephanou, Environ. Sci. Technol. 41 (2007)
8011.
[43] C. Schauer, R. Niessner, U. Pöschl, Environ. Sci. Technol. 37
(2003) 2861.
[44] European Commission, Directive 2004/107/EC of the European
Parliament and of the Council of 15 December 2004 relating to
arsenic, cadmium, mercury, nickel and polycyclic aromatic
hydrocarbons in ambient air, Offic. J. Eur. Un. 2005, L23, 3–16.
[45] E. Menichini, Chemosphere 77 (2009) 1275.
[46] H. Creutznacher, M. Cipolloni, F. Loffler, Gefahrstoffe Reinhaltung
Der Luft 65 (2010) 437.
[47] M. Possanzini, V. Di Palo, P. Gigliucci, M.C. Tomasi Sciano, A.
Cecinato, Atmos. Environ. 38 (2004) 1727.
[48] W.-J. Lee, Y.-F. Wang, T.-C. Lin, Y.-Y. Chen, W.-C. Lin, C.-C. Ku,
J.-T. Cheng, Sci. Total Environ. 159 (1995) 185.
[49] E. Wauters, P. Van Caeter, G. Desmet, F. David, C. Devos, P.
Sandra, J. Chromatogr., A 1190 (2008) 286.
[50] Z.H. Fan, K.H. Jung, P.J. Lioy, Environ. Sci. Technol. 40 (2006)
6051.
[51] C.C. Cheng, Polycycl. Aromat. Compd. 23 (2003) 315.
[52] A. Wnorowski, M. Tardif, D. Harnish, G. Poole, C.H. Chiu,
Polycycl. Aromat. Compd. 26 (2006) 313.
Trends
[53] A. Christensen, C. Ostman, R. Westerholm, Anal. Bioanal. Chem.
381 (2005) 1206.
[54] M. Ryno, L. Rantanen, E. Papaioannou, A.G. Konstandopoulos, T.
Koskentalo, K. Savela, J. Environ. Monit. 8 (2006) 488.
[55] S.C. Zhang, W. Zhang, K.Y. Wang, Y.T. Shen, L.W. Hu, X.J.
Wang, Environ. Monit. Assess. 151 (2009) 197.
[56] European Committee for Standardization (CEN), EN 15549:2008
Air Quality- Standard method for the measurement of the
concentration of benzo[a]pyrene in air, CEN, Brussels, Belgium,
2008.
[57] A. Albinet, E. Leoz-Garziandia, H. Budzinski, E. ViIlenave, J.
Chromatogr., A 1121 (2006) 106.
[58] M.F. Simcik, H. Zhang, S.J. Eisenreich, T.P. Franz, Environ. Sci.
Technol. 31 (1997) 2141.
[59] B.L. van Drooge, I. Nikolova, P.P. Ballesta, J. Chromatogr., A
1216 (2009) 4030.
[60] M. Bates, P. Bruno, M. Caputi, M. Caselli, G. de Gennaro, M.
Tutino, Atmos. Environ. 42 (2008) 6144.
[61] J. Gil-Molto, M. Varea, N. Galindo, J. Crespo, J. Chromatogr., A
1216 (2009) 1285.
[62] P.B.C. Forbes, E.R. Rohwer, Environ. Pollut. 157 (2009) 2529.
[63] S.K. Pandey, K.-H. Kim, Environ. Sci. Technol. 43 (2009)
3020.
[64] D.Z. Bezabeh, H.A. Bamford, M.M. Schantz, S.A. Wise, Anal.
Bioanal. Chem. 375 (2003) 381.
[65] L.-C. Ding, F. Ke, D.K.W. Wang, T. Dann, C.C. Austin, Atmos.
Environ. 43 (2009) 4894.
[66] D. Waterman, B. Horsfield, F. Leistner, K. Hall, S. Smith, Anal.
Chem. 72 (2000) 3563.
[67] J.M. Vaz, Talanta 60 (2003) 687.
[68] E. Baltussen, H.-G. Janssen, P. Sandra, C.A. Cramers, J. High Res.
Chromatogr. 20 (1997) 385.
[69] R.J.C. Brown, R.E. Yardley, A.S. Brown, P.R. Edwards, C. Rivier, C.
Yardin, Anal. Lett. 39 (2006) 1229.
[70] R.J.C. Brown, Sci. World J. 8 (2008) 796.
[71] A.S. Brown, R.J.C. Brown, J. Automat. Methods Manage. Chem.,
Article ID 179498 (2008) 14 pp. (doi: 10.1155/2008/179498).
http://www.elsevier.com/locate/trac 1739