Trends Trends in Analytical Chemistry, Vol.

46, 2013

Worldwide distribution of persistent

organic pollutants in air, including
results of air monitoring by passive
air sampling in five continents
C. Bogdal, E. Abad, M. Abalos, B. van Bavel, J. Hagberg, M. Scheringer,
H. Fiedler
This article provides an overview of concentrations of persistent organic pollutants (POPs) in ambient air on a global scale,
including recent measurements and an extensive compilation of literature data. In this study, passive air samplers (PASs) were
successfully employed to assess concentrations of POPs in ambient air from Africa, Latin America, the Caribbean, and the Pacific
Islands. The project aimed to extend the knowledge on environmental contamination by POPs in these regions, where the
currently available data are still limited to a few monitoring studies.
The ambient air concentrations of PCB in Africa were relatively high when compared to other regions. Waste, in particular
electronic waste, exported to Africa from industrialized countries may be a possible source of PCB in Africa, where PCB have
never been extensively used or produced. For DDTs, the wide range of concentrations and particularly high levels in some
countries of Africa and the Pacific Islands reflect the use of DDT for malaria control in these regions. For PCDD/PCDF,
concentrations in Africa and Latin America are similar to or even higher than in Europe, probably due to unfavorable combustion
practices of chlorine-containing materials.
The data support the needs for further monitoring in developing countries and countries with economies in transition, and
action to reduce environmental contamination by, and human exposure to, hazardous chemicals.
ª 2012 Elsevier Ltd. All rights reserved.

Keywords: Air monitoring; Dichlorodiphenyl trichloroethane (DDT); Dioxin-like PCB (dl-PCB); Environmental contamination; Human exposure;
Passive air sampling; Persistent organic pollutant (POP); Polychlorinated biphenyl (PCB); Polychlorinated dibenzo-p-dioxin (PCDD); Polychlorinated
dibenzofuran (PCDF)

C. Bogdal*, M. Scheringer
1. Introduction
Institute for Chemical and Bioengineering, ETH Zurich, Wolfgang-Pauli-Strasse 10, CH-8093 Zürich,
Switzerland Persistent organic pollutants (POPs) rep-
International Panel on Chemical Pollution, Office ETH Zurich, Wolfgang-Pauli-Strasse 10, CH-8093 resent a group of organic chemicals of
Zürich, Switzerland particular concern because they are
E. Abad, M. Abalos
resistant to degradation in the environ-
Laboratory of Dioxins, Department of Environmental Chemistry, IDÆA-CSIC, Jordi Girona 18, ment (persistent), accumulate in human
08034 Barcelona, Spain and animal tissue, biomagnify along food
chains, and are associated with chronic
B. van Bavel, J. Hagberg (eco)toxicological effects [1]. In 2004, the
Man-Technology-Environment (MTM) Research Centre, Örebro University, SE-701 82 Örebro,
Sweden
Stockholm Convention on POPs came into
force and aims to protect human health
H. Fiedler and the environment from negative effects
UNEP Chemicals, 11-13, Chemin des Anémones, CH-1219 Chaˆtelaine (GE), Switzerland of POPs by reducing or eliminating re-
leases of POPs to the environment [2]. In
*
2009, the initial list of 12 POPs was
Corresponding author.
Tel.: +41 44 632 5951;
extended by nine additional chemicals.
E-mail: christian.bogdal@chem. POPs have become ubiquitous contam-
ethz.ch inants found in not only the environment

150 0165-9936/$ - see front matter ª 2012 Elsevier Ltd. All rights reserved. doi:http://dx.doi.org/10.1016/j.trac.2012.05.011
Trends in Analytical Chemistry, Vol. 46, 2013 Trends

and biota of their region of initial use, but also remote 2. Determination of levels of POPs in ambient air
areas far away from their primary emission sources [3–
7]. POPs used as pesticides are directly emitted into the To sample air volumes that are large enough to over-
atmosphere by spraying application. A considerable come analytical detection limits, PASs have been devel-
amount of POPs is also released to air by volatilization oped as simple, cost-effective alternatives to active air
from POP-containing applications (open sources) or in samplers [24]. The core part of a PAS consists of an
fumes from combustion processes (unintentionally adsorbent that is exposed to a passive air flow. Airborne
formed POPs), and the atmosphere is a key medium for chemicals are taken up by molecular diffusion, resulting
environmental long-range transport of POPs, next to the in an accumulation of chemicals in the adsorbent. In the
slower transport by rivers, ocean currents, and sea ice. initial so-called ‘‘linear uptake phase’’, the amount of
Ambient air monitoring of POPs is therefore particularly airborne chemical sequestered by the adsorbent in-
relevant for investigating environmental levels and creases linearly with time and – under the assumptions
trends of POPs. Air has also been recognized as a key of a constant volumetric sampling rate of air and a
environmental medium in the Global Monitoring Plan of constant chemical concentration in air – is proportional
the Stockholm Convention, which is intended to support to the sampled air volume (Fig. 1a). In the course of
evaluation of the effectiveness of the Convention. In exposure time, the uptake rate decreases until equilib-
addition, several monitoring studies on POPs in the rium between the adsorbent and ambient air is reached.
atmosphere on a local, regional, and global scale have Thus, when PASs are deployed for a period not longer
been conducted [8–17]. than the linear uptake phase, the amount of chemical
Still, ambient air monitoring of POPs is a challenging sequestered in the adsorbent makes it possible to deter-
task. On the one hand, concentrations of atmospheric mine the concentration of the pollutant in ambient air,
contaminants are variable spatially and, in particular, based on the sampled volume of air.
temporarily. On the other hand, the usually low con- Polyurethane foam (PUF) disks proved to be an ade-
centrations of POPs in air inevitably require elaborate quate adsorbent in PASs [24]. PASs used in the present
sampling techniques followed by high-performance, re- project are identical to the devices used in several pre-
source-intensive analytical methods, so there are only a vious studies (Fig. 1b) [14,16,24] and consist of a PUF
few large monitoring studies about POPs in air and the disk protected from dry and wet deposition by a stainless-
global coverage is still unequal. Whereas fairly extended steel casing, which has also a wind-shielding function to
data sets about atmospheric occurrence of POPs are reduce fluctuations of the air flow [25,26].
available for the Northern hemisphere [8,11,12,18–22]; Low acquisition costs, limited maintenance demand,
Africa [14,16], Latin America [16,17], and the Pacific simple handling, and no need for power supply are
region [16] have scarcely been covered. advantages offered by PASs, especially in countries with
Herein, we present a recently conducted study about limited financial resources or in remote areas. However,
POPs in air in (i) West, East, and South Africa (ii) Latin the accurate determination of ambient-air concentra-
America and the Caribbean, and (iii) Pacific Islands. We tions represents a difficulty when PASs are used. The
focus on polychlorinated biphenyls (PCBs), dichlorodi- analysis of pollutants in a PUF provides merely their
phenyl trichloroethane (DDT) and its main transforma- amount sequestered in the PUF during the deployment
tion products (together referred to as DDTs), as well as period. Only after dividing the amount of chemicals (e.g.,
polychlorinated dibenzo-p-dioxins and dibenzofurans expressed in pg) by the corresponding air-sampling vol-
(PCDDs/PCDFs), as representative of industrial POP ume (e.g., expressed in m3) is the concentration of POPs
chemicals, POP pesticides, and unintentionally formed in air obtained (e.g., expressed in pg/m3). The determi-
POPs, respectively. nation of the sampling volume requires extensive cali-
DDTs are of particular interest with respect to the bration exercises. Uptake rates of POPs (e.g., PCB) can
current exemption under the Stockholm Convention, vary from 2–8 m3/day [24,27] and typically lie between
permitting the use of DDT for malaria control in af- 3–4 m3/day [14,16,24,27]. The deployment period has
fected regions. Extensive results for 12 POPs measured to be long enough to obtain the required air-sampling
in ambient air samples, including individual PCB, DDT, volume, but it should not be too long to avoid the PUF
and PCDD/PCDF congeners, are presented in a com- reaching equilibrium with ambient air (Fig. 1a). Typical
panion document [23]. First, we introduce the applied sampling periods lie within a few weeks, resulting in
method based on passive air samplers (PASs). Second, sampled air volumes of 100–500 m3.
we present and discuss the collected data set in a wider For POPs, another limitation needs to be considered.
context, including comparison to existing monitoring Because of their low vapor pressure, POPs in air partition
studies, for which the raw measurement data are between the pure gas phase and the particulate phase
available and which rely on the same sampling tech- (aerosols). Uptake rates determined so far apply to only
nique. the gas phase. The entrapment of aerosols and associated

http://www.elsevier.com/locate/trac 151
Trends Trends in Analytical Chemistry, Vol. 46, 2013

sequestred chemical
equilibrium
amount of phase

e ke
a s ta
ph r up
ea
lin

time

casing target chemical

PUF

10 cm
air flow
Figure 1. (a) Uptake phases of airborne chemicals by passive air samplers (adapted from [24]); and, (b) polyurethane foam based passive air
sampler (adapted from [12]). Note that in Fig. 1(b), the symbols for the molecules of the target chemical are not true to scale.

POPs by PUFs installed in PASs is hardly known. In  11 countries in West, East, and South Africa (The
2009, Chaemfa and co-workers reported on the uptake Democratic Republic of the Congo, COD; Ethiopia,
of aerosols by PASs and mentioned that ‘‘some’’ sam- ETH; Ghana, GHA; Kenya, KEN; Mali, MLI; Mauritius,
pling of particles and associated chemicals occurs, but a MUS; Nigeria, NGA; Senegal, SEN; Togo, TGO; Ugan-
quantitative determination of uptake rates was not da, UGA; Zambia, ZMB);
provided [25]. In the present study, this limitation is  12 countries in Latin America and the Caribbean
worth noting for the interpretation of data on PCDDs/ (Antigua and Barbuda, ATG; Bahamas, BHS; Brazil,
PCDFs. For PCDDs/PCDFs (logarithmic octanol-air par- BRA; Chile, CHL; Cuba, CUB; Ecuador, ECU; Haiti,
tition coefficient at 25C, logKoa around 10–12), the HTI; Jamaica, JAM; Mexico, MEX; Peru, PER; Uru-
fraction associated with atmospheric particles is higher guay, URY); and,
than for PCB (logKoa around 8–10) and DDTs (logKoa  seven countries in the Pacific Region (Fiji, FJI; Kiribati,
around 9–10) [28]. KIR; Niue, NIU; Palau, PLW; Samoa, WSM; Solomon
In this project, ambient air was sampled in 31 coun- Islands, SLB; Tuvalu, TUV).
tries with PASs equipped with PUF disks (Fig. 2). Details PASs were deployed for three months in Africa and
of the sampling procedure are presented in an accom- Latin America, and three or six months in the Pacific
panying UNEP report [23]. The three regions covered Islands. After recovery of PUFs from PASs after exposure,
include: POPs were measured in PUFs in the laboratories of:

152 http://www.elsevier.com/locate/trac
Trends in Analytical Chemistry, Vol. 46, 2013 Trends

 IVM Free University Amsterdam (VU University) (indi- 3. Worldwide distribution of POPs in ambient air
cator PCBs and organochlorine pesticides, including
DDTs in samples from Africa and the Pacific Islands); 3.1. Polychlorinated biphenyls (PCBs)
 MTM Research Center, Örebro University (dioxin-like The distribution of PCBs in ambient air observed in this
PCBs and PCDDs/PCDFs in samples from Africa and study is shown in Fig. 3. Our data set is compared to
the Pacific Islands); and, further monitoring and research studies relying on the
 IDÆA-CSIC, Barcelona (all POPs in samples from La- same PAS technique in Fig. 4, where ranges of concen-
tin America and the Caribbean). trations of PCBs in ambient air are grouped in five geo-
Analysis of POPs in PUFs was based on surrogate graphical regions, including Africa, Pacific and
standard addition, Soxhlet extraction, and further Australia, Asia, Europe, Latin America, and North
purification of extracts. Measurements were performed America. A compilation of worldwide measurements of
with gas chromatography (GC) electron-capture detec- PCBs, DDTs, PCDDs/PCDFs, and dl-PCB, including the
tor (ECD) at IVM VU University for POPs pesticides and data from this study, is provided in Table S1 in the
indicator PCB, and GC coupled to high-resolution mass Supplementary Material.
spectrometry (HRMS) at MTM and IDÆA-CSIC. To In this study, the median concentration of R7PCB in
facilitate comparison of our results with existing data, ambient air was highest in the samples from Africa
results for the sequestered amount of POPs in PASs (median 84 pg/m3), followed by the Pacific Islands
were converted into concentrations based on the (24 pg/m3), and Latin America (21 pg/m3) (Fig. 3). The
deployment time of PASs and a typical sampling rate of range of concentrations in this study (min–max: 1–
3.5 m3/d. Literature data were also converted into 7281 pg/m3) was highest in the Latin America region,
concentrations, if necessary, based on the information where the urban site in La Habana, Cuba (CUB), clearly
about sampling time and rate provided. Regarding the has the highest concentration. Also samples from Africa
principle of operation of PASs, as well as their limita- had a wide range (8–2074 pg/m3), particularly with the
tions, the literature data was mainly restricted to urban site in Bamako, Mali (MLI), having the highest
comparable studies relying on PASs. Whenever possible, concentration. Samples from the Pacific Islands were all in
the sum of the seven indicator PCBs [2,4,4 0 ,trichloro- a relatively narrow concentration range (4–108 pg/m3).
biphenyl (PCB 28), 2,2 0 ,5,5 0 -tetrachlorobiphenyl (PCB The congener pattern of PCBs remained fairly con-
52), 2,2 0 ,4,5,5 0 -pentachlorobiphenyl (PCB 101), 2,3 0 ,4, stant for all samples in the three regions. Among PCBs,
4 0 ,5-pentachlorobiphenyl (PCB 118), 2,2 0 ,3,4,4 0 , the lower chlorinated congeners dominate (PCB
5 0 -hexachlorobiphenyl (PCB 138), 2,2 0 ,4,4 0 ,5,5 0 - 28  PCB 52), followed by the higher chlorinated
hexachlorobiphenyl (PCB 153), and 2,2 0 ,3,4,4 0 ,5,5 0 - congeners (PCB 101 P PCB 118  PCB 138  PCB
heptachlorobiphenyl (PCB 180)] (denoted hereafter 153 > PCB 180).
R7PCB) and the sum of op 0 - and pp 0 -congeners of DDT, Concentrations measured in this study are similar to
DDD (dichlorodiphenyl dichloroethane), and DDE those measured in previous monitoring studies (Fig. 4).
(dichlorodiphenyl dichloroethene) (denoted hereafter In Africa, large data sets are available from the MONET
DDTs) were considered (with the exception of some monitoring network (R7PCB, median: 52 pg/m3; min–
studies that considered different analytes). All 17 max: 2–1324 pg/m3) [14] and a study performed re-
2,3,7,8-chlorosubstituted PCDDs/PCDFs (denoted here- cently in West Africa (R7PCB: 17; 3–125 pg/m3) [9]. For
after PCDDs/PCDFs) were considered, and concentra- the Pacific region, the samples from Australia measured
tions were converted into TEQ based on the WHO98- in the GAPS study (R48PCB: 7 and 13 pg/m3) [16] are
TEFs [29], which is the scheme recommended according within the range found in this study, although the GAPS
to the Stockholm Convention on POPs (Annex C of the study included 48 PCB congeners and, therefore, prob-
Stockholm Convention) [2]. Also, dioxin-like PCBs ably provided higher concentrations. In Latin America,
[including 3,3 0 ,4,4 0 -tetrachlorobiphenyl (PCB 77), the concentrations reported for Chile (R48PCB: 13; 2–
3,4,4 0 ,5-tetrachlorobiphenyl (PCB 81), 2,3,3 0 ,4,4 0 -pen- 53 pg/m3) [17] and for different countries in the GAPS
tachlorobiphenyl (PCB 105), 2,3,4,4 0 ,5-pentachlorobi- study (R48PCB: 10; 1–32 pg/m3) [16] are similar to
phenyl (PCB 114), 2,3 0 ,4,4 0 ,5-pentachlorobiphenyl those in this study.
(PCB 118), 2,3 0 ,4, 4 0 ,5 0 -pentachlorobiphenyl (PCB For comparison, Fig. 4 includes PCB concentration
123), 3,3 0 ,4,4 0 ,5-pentachlorobiphenyl (PCB 126), ranges in Asia, Europe, and North America. In Asia, the
2,3 0 ,4,4 0 ,5,5 0 -hexachlorobiphenyl (PCB 167), 2,3, range of reported concentrations is remarkably large
3 0 ,4,4 0 ,5-hexachlorobiphenyl (PCB 156), 2,3,3 0 ,4,4 0 ,5 0 - with Jaward and co-workers (2005) [13] reporting the
hexachlorobiphenyl (PCB 157), 3,3 0 ,4,4 0 ,5,5 0 -hexa- lowest concentrations for samples from China, Japan,
chlorobiphenyl (PCB 169), and 2,3,3 0 ,4,4 0 ,5,5 0 -hepta- South Korea, and Singapore (R7PCB: 13; 1–115 pg/m3)
chlorobiphenyl (PCB 189)] (denoted hereafter dl-PCBs) and Pozo et al. (2011) [15] reporting the highest
were measured and converted into WHO98 TEQ. concentrations in agricultural sites in India (R48PCB:

http://www.elsevier.com/locate/trac 153
Trends Trends in Analytical Chemistry, Vol. 46, 2013

Figure 2. Sampling sites included in the present study, covering Latin America and the Caribbean; West, East, and South Africa; and the Pacific
region. Countries are labeled by their ISO 3166-1 alpha-3 code, further defined in the text.

290; 20–334000 pg/m3). In Europe, concentrations of 3.2. Dichlorodiphenyl trichloroethane and

PCBs measured at background sites in the EMEP moni- transformation products (DDTs)
toring program are among the lowest concentrations A compilation of concentrations of DDTs in ambient air
observed (R7PCB: 17; 2–121 pg/m3) [11], whereas, in measured in this study is provided in Fig. 3 and a
mixed urban/rural areas in Europe (R48PCB: 68; 6– comparison between the five regions is shown in Fig. 5.
644 pg/m3 by Pozo et al. (2009) [16], and R7PCB: 17; In this study, concentrations of DDTs in ambient air was
2–121 pg/m3 by Jaward et al. (2004) [12]) and in the highest in the Pacific Islands (median: 306 pg/m3) and
urban area of Philadelphia, USA, higher concentrations in Africa (51 pg/m3), and much lower in Latin America
were reported (R5PCB without PCB 118/180: 63; 15– (<1 pg/m3). Also the range of concentrations was largest
849 pg/m3) [8]. in the Pacific Islands (min–max: 27–2255 pg/m3) with
Only a few studies applying active air sampling are highest concentrations in Samoa (WSM) and Solomon
available for Africa and provide results comparable to Islands (SLB), as well as in Africa (8–2178 pg/m3) with a
our measurements. In the MONET monitoring network, high concentration in Mali (MLI). In Latin America, the
active air sampling was performed at two sites in Uganda concentration range was much lower and much nar-
in 1999–2006 and showed concentrations of R7PCB of rower (<1–5 pg/m3).
30–1729 pg/m3, overlapping with the range found in Concentrations measured here for Africa are still in
this study for Uganda (84–101 pg/m3) [14]. During a the same range as reported in the MONET study for most
ship cruise in 2007 along the west African coast, Gioia sites (median: 134 pg/m3; min–max: 5–8000 pg/m3),
and co-workers measured concentrations of 16–272 pg/ apart from the sample from an obsolete pesticide-
m3 for R7PCB [9], which is comparable to our concen- dumping site, where concentrations up to 90,000 pg/m3
trations observed in Senegal (SEN: 8–42 pg/m3), Ghana have been reported [14]. Active air sampling performed
(GHA: 38–74 pg/m3), and Togo (TGO: 8–33 pg/m3). in Uganda within the MONET project revealed

154 http://www.elsevier.com/locate/trac
Trends in Analytical Chemistry, Vol. 46, 2013 Trends

Figure 3. Distribution of R7PCB and DDTs in Latin America and the Caribbean; West, East, and South Africa; and the Pacific region observed in
this study. Countries are labeled by their ISO 3166-1 alpha-3 code, further defined in the text.

concentrations of 22–733 pg/m3, going up higher than the concentrations reported by Jaward et al. (2005) [13]
in our study (22–37 pg/m3), as already observed for for China, Japan, South Korea, and Singapore
PCBs. The few samples from Latin America analyzed (<1–1715 pg/m3) and Pozo et al. (2011) [15] for Indian
within the GAPS study [16] showed concentrations of agricultural sites (3–10,930 pg/m3) are similar to the
DDT (<1–4 pg/m3) very similar to this study. In Asia, African sites. In Europe, however, concentrations of

http://www.elsevier.com/locate/trac 155
Trends Trends in Analytical Chemistry, Vol. 46, 2013

Africa Pacific & Asia 334000 Europe Latin America North

10000 Australia
291000
90600 America

PCB concentration in air [pg/m3]

Philippines
DR Congo, Egypt,

Cuba
Mali, Senegal
Mali
1000

100

10

1
n = 156
n = 44

n = 15

n = 23

n = 77

n = 56

n = 71

n = 83

n = 32
n= 62
n=2

n=2

n=6

n=8

n=6

n=6

n=6
0.1
y)

)
)

S)
y)
S)

)
)

5)

4)

S)

S)

)
4)
y)
11
ET

PS

1
PS

EP

09
ud

ud

01

AP
00

ud
00
AP

00
AP

AP
20
N

20
st

A
A

M
st

2
.2

st
2

2
O

(G
(G

(G
(G

(G

(G
(E
ia
is

u
o
is
(M

is
al

oz
io

ar
(th

(D
oz
e
(th
a

ia
lia

an

PA rica
(th
nd
(G

p
et
ric

w
(P
a

As

(P
ro
tra

SA
a

be
ou
ric

a
s

n
Af

e
ric

nd

Eu

(J
a

e
ea
ar
us

m
ib
Af

,U
ric

di

l
g
Af

hi
la

e
aw

ar

A
b
ck
A

In
Af

op

C
Is

ib

th
ba
(J

Am Car
r

or
fic

&
Eu

a,
ia

pe

N
ci

hi
As

&

ic
Pa

ro

lp
er
a

de
Eu

ic
er

ila
Am

Ph
tin
La
tin
La
Figure 4. Concentrations of PCB based on passive air sampling in different global monitoring studies. This figure is a boxplot with the box stretch-
ing from the 25th percentile to the 75th percentile and containing the median in between. The whiskers indicate the minimum and maximum
observed value or 1.5 times the interquartile range. When there are outliers lower or higher than 1.5 times the interquartile range, these are rep-
resented separately by circles. Where the number of data points is too low to be displayed as box plot, the data points are represented by tri-
angles. Note the logarithmic scale on the Y axis. The references of the original studies and comments are provided in Table S1.
Pacific &
Australia

Africa Asia Europe Latin North

100000
DDTs concentration in air [pg/m3]

America America
Kenya

10000

1000

100

10

1
n = 156
n = 44

n = 23

n = 77

n = 55

n = 69

n = 86

n = 62

0.1
n=1

n=3

n=8

n=3

n=1

0.01
)

S)

)
y)
y

S)

(G )
ET

S)
y)
PS
11
05

04

(th EP

PS
ud

ud
AP

ud
P

AP
N

20
20

20
A
st

A
st
O

st
(G

(G

(G

(G
be d (E
is

Eu zo
(M

d
is

is
(th

ar

ar
a

pe
(th

In sia

an

a
o
ric
a

aw

aw

ic
(P

un
ro
a

ric

be
A

er
s
Af

an
ric

ro
nd

(J

J
a
Af

Am
ib
(
di
Af

g
la

ia

pe

ar
ck
Is

As

ib

th
C
ro

ba

ar

or
c

&
Eu
fi

C
e

N
ci

a
Am rop

&

ic
Pa

er
a
Eu

ic

Am
er
tin
La
tin
La

Figure 5. Concentrations of DDTs (including o,p 0 - and p,p 0 - congeners of DDT, DDD, and DDE, unless specified) based on passive air sampling
in different global monitoring studies. This figure is a boxplot with the same representation as in Fig. 1. Where the number of data points is too
low to be displayed as box plot, the data points are represented by triangles. Note the logarithmic scale on the Y axis. The references of the
original studies and comments are provided in Table S1.

DDTs appear to be lower, as reported in the GAPS 3.3. Polychlorinated dibenzo-p-dioxins and dibenzofu-
monitoring study (9–192 pg/m3) [16], in the study by rans (PCDDs/PCDFs) and dioxin-like PCBs (dl-PCBs)
Jaward et al. (2004) (2–311 pg/m3) [12], and at the Fig. 6 shows the global distribution of PCDDs/PCDFs and
background sites in the EMEP network (1–356 pg/m3) dl-PCB in ambient air observed in this study, while Fig. 7
[11]. provides a comparison to other studies. Concentrations

156 http://www.elsevier.com/locate/trac
Trends in Analytical Chemistry, Vol. 46, 2013 Trends

Figure 6. Distribution of PCDD/PCDF and dl-PCB in Latin America and the Caribbean; West, East, and South Africa; and the Pacific region
observed in this study. Countries are labeled by their ISO 3166-1 alpha-3 code, further defined in the text.

of PCDDs/PCDFs in this study were highest in air sam- bana, Cuba (CUB) (2042 fg WHO98 TEQ/m3). Dl-PCBs
ples from Latin America (median: 74 fg WHO98 TEQ/m3; were lowest in the Pacific Islands (only three samples
min–max: 9–678 fg WHO98 TEQ/m3) and Africa (51; available: 3, 13, and 23 fg WHO98 TEQ/m3). In Africa,
18–532 fg WHO98 TEQ/m3). In the Pacific Islands, the median contribution of dl-PCBs to the total TEQ is
concentrations were lower by an order of magnitude (3; 15% (range: 5% (Mauritius) to 60% (Zambia)). In Latin
1–87 fg WHO98 TEQ/m3). Concentrations of dl-PCBs America, the median dl-PCB contribution to the total
were also highest in Africa (11; 7–278 fg WHO98 TEQ/ TEQ is 23% (range: 4% (Mexico) to 84% (Cuba)). For dl-
m3). In Latin America, dl-PCBs had concentrations PCB and in comparison with other data from Cuba and
similar to Africa (14; 1–248 fg WHO98 TEQ/m3), except elsewhere in Latin America, the site in La Habana must
for the much higher concentration measured in La Ha- represent a particularly exposed site rather than a

http://www.elsevier.com/locate/trac 157
Trends Trends in Analytical Chemistry, Vol. 46, 2013

baseline national site. Such conclusion also applies to the have been reported off the West African coast during a
indicator PCB. ship cruise in the Atlantic Ocean in 2005 [10]. Recently,
PCDDs/PCDFs were also measured by PASs in the another ship cruise along the West African coast com-
MONET monitoring study in Africa [14]. There, median bined with passive air sampling in West African coun-
concentrations were similar to the current study (55 fg tries (Gambia, Ghana, Ivory Coast, and Sierra Leone)
WHO98 TEQ/m3), although the two samples from Egypt confirmed the high continental emissions of PCBs [9].
had considerably higher concentrations (5464 and Also, the elevated concentrations of PCB and the
6548 fg WHO98 TEQ/m3). A few individual measure- increasing trend between 2004 and 2009 in human
ments of PCDDs/PCDFs from active air sampling are breast milk from Ghana was related to human exposure
available for Latin America. De Assunção et al. found to toxic waste from industrial countries [37].
concentrations of 67–766 fg WHO98 TEQ/m3 in São Another important source of PCBs in developing
Paulo, Brazil, in 2000–2001 [30]. More recently, Card- countries might be ship-dismantling activities (also re-
enas et al. reported a concentration range of <1–410 fg ferred to as ship breaking or ship wreckage). Old ships
WHO05 TEQ/m3 (data reported using the WHO 2005 containing PCBs in paints, transformer and capacitor
TEF values) for rural and urban sites in Mexico in 2008– oils, and sealants, are regularly exported to developing
2010 [31], while, in 2009-2010, Aristoza´bal et al. countries or countries with economy in transition,
measured 1–52 fg WHO98 TEQ/m3 in particulate matter where they are dismantled under primitive conditions
in Manizales, an Andean city in Colombia [32]. These [38,39]. India is one of the most important ship-
active air-sampling measurements are in the same range dismantling countries in Asia and worldwide [40]. Pozo
as in our study for Brazil (37–167 fg WHO98 TEQ/m3) et al. (2011) [15] also related the high concentrations of
and Mexico (22 fg WHO98 TEQ/m3). PCBs in air found in India to e-waste recycling activities
and/or open municipal waste-dumping sites.
Also, in Africa, the existence of numerous abandoned
4. Discussion ships on the coast, particularly in West Africa, is known
(http://www.shipbreakingplatform.org) (e.g., Nouadhi-
4.1. Polychlorinated biphenyls (PCBs) bou, the second largest city in Mauritania, has been re-
Global production of PCBs peaked in the early 1970s and ported to host one of the biggest ship cemeteries in the
ceased in the early 1990s. The main production and use world).
of PCBs occurred in North America, Europe, the former Weber and co-workers (2008) [41], and more recently
Soviet Union, and Japan [33]. Considering that PCB Breivik et al. (2011) [42] extensively discussed the issue
production ended more than 20 years ago, concentra- of toxic waste exports from industrialized countries to
tions of PCBs in air in some regions are still remarkably developing countries and countries with economies in
high. In Africa and South-East Asia, where PCBs were transition. E-waste handling and ship-breaking activities
never extensively produced and used, concentrations in in developing countries have been identified as major
air are similar to or even higher than in former pro- environmental issues, not only because these practices
duction and use regions of PCBs (e.g., Europe and North result in potential environmental emissions, but also
America). Thus, it appears that, in these regions, PCB- because these emissions are further enhanced by cli-
containing products are present and emissions continue. matic conditions. The warm conditions prevailing in
Possible sources of PCBs in Africa and Asia may be re- tropical regions increase volatilization rates of semi-
lated to handling of waste of electrical and electronic volatile chemicals, because the vapor pressure increases
equipment (e-waste), open waste dumping, and ship- exponentially with temperature.
dismantling operations. The PASs from La Habana, Cuba
(CUB), were positioned in the harbor with intense com- 4.2. Dichlorodiphenyl trichloroethane and transfor-
mercial ship traffic. Furthermore, in the vicinity of the mation products (DDTs)
sampling site (1–2 km aerial distance), a large petrol DDT is a particular POP pesticide, because it is still used in
refinery is located. These two observations may explain some countries as an insecticide in the public health
the remarkably high concentrations of PCBs observed at sector for malaria control. In 2006, an UNEP expert
this site. group assessed the contemporary DDT production and
The current global production of e-waste is estimated use and its alternatives [43]. The total global production
to be 20–25 million t/a, with most e-waste being pro- for DDT for vector control was estimated at 6269 t/a in
duced in Europe, USA and Australasia [34]. E-waste 2006, with the production occurring mainly in India and
handling and open waste dumping has been found to be China. There were 21 countries in Africa, Asia, Pacific,
a relevant source of PCBs in China [35–36]. and the Middle East with known or potential DDT use for
Also, in Africa, indications for relevant emissions of disease vector control. In Africa, Cameroon, Malawi,
PCBs, possibly due to waste handling, have been pro- Nigeria, South Africa, Tanzania, and Uganda use DDT for
vided. Elevated concentrations of PCB in ambient air malaria control. Ethiopia, Mozambique, Zambia and

158 http://www.elsevier.com/locate/trac
Trends in Analytical Chemistry, Vol. 46, 2013 Trends

concentration in air [fg WHO98 TEQ/m3]

Africa Pacific Latin America

) and dl-PCB ( )
10000 Islands

Egypt

Cuba
1000

100

10
PCDD/PCDF (
1

n = 26

n = 27

n = 27
n=9

n=9

n=8

n=3
0.1

y)

)
)

y)
y)
dy

dy

dy
ET

ud

ud
ud
u

tu
N
st

st

st

st
st

s
O
is

is

is
is
(M

i
(th

(th

(th

(th

(th
(th
a
a

s
ric

an

an
s

nd
ric

ric

d
Af

be
n

be
Af

Af

la
la

ib
Is
Is

i
ar

ar
fic
fic

C
ci
ci

&

&
Pa
Pa

a
ic

ic
er

er
Am

Am
tin

tin
La

La

Figure 7. Concentrations of PCDD/PCDF (including all 17 2,3,7,8-chloro substituted PCDD/PCDF) and dl-PCB (including PCB 77/81/105/114/
118/123/126/156/157/167/169/189) based on passive air sampling in different global monitoring studies. This figure is a boxplot with the same
representation as in Fig. 1. Where the number of data points is too low to be displayed as box plot, the data points are represented by triangles.
Note the logarithmic scale on the Y axis. Continuous lines represent data for PCDD/PCDF; dotted lines represent data for dl-PCB. The references
of the original studies and comments are provided in Table S1.

Zimbabwe reported increases in DDT use since 2000. which is in agreement with the GAPS study [16] and
Angola, Botswana, Ecuador, the Philippines, and Senegal reflects the phasing out of DDT in this region.
reported keeping stocks in the event of malaria outbreaks
in those countries. Besides use in the Dominican Republic 4.3. Polychlorinated dibenzo-p-dioxins and dibenzofu-
for disease vector control, no other use of DDT has been rans (PCDDs/PCDFs) and dioxin-like PCBs (dl-PCBs)
reported from the rest of the Americas, with Ecuador, Due to the demanding analytical techniques required for
Mexico and Venezuela being the last countries to have the determination of PCDDs/PCDFs, only a few data sets
phased out DDT in 2000. are available for ambient air, particularly with passive air
The large range of concentrations in ambient air sampling. Concentrations of PCDDs/PCDFs in Africa and
within each region and the differences between regions Latin America measured in this project are similar to
reflect the different patterns of use and regulation. The concentrations reported for Europe based on active air
peak concentrations observed in Africa are probably a sampling.
result of DDT use in these regions. In this project, highest In urban, industrial, and rural sites in Catalonia,
concentrations were observed in Zambia (ZMB) and Mali Spain, Abad et al. (2007) [18] reported concentrations of
(MLI). In the MONET project, Kenya had the highest PCDDs/PCDFs of 5–1196 fg I-TEQ/m3. For an urban site
concentrations, followed by Senegal, Sudan, Mali, and in Rome, 5–734 fg WHO98 TEQ/m3 were reported [21].
Ethiopia [14]. Recently, Umlauf et al. (2010) [22] reported concen-
In this project, relatively high concentrations of DDTs trations of 40–3200 fg WHO98 TEQ/m3 in urban,
were measured in the Pacific Islands, where DDT is industrial, and rural sites in the Malopolska Region,
sprayed for malaria control [44,45]. Other studies also southern Poland. Katsoyiannis et al. (2010) [20] re-
showed high concentrations and broad ranges of DDTs ported concentrations of <50 fg WHO98 TEQ/m3 for
in Asia. By contrast, at European background sites major urban areas in the UK.
where DDT was banned several decades ago, concen- Considering that PASs do not sample PCDDs/PCDFs
trations are much lower. Interestingly, in Latin America, with complete efficiency due to the incomplete sampling
the lowest concentrations were observed in this project, of the atmospheric particulate fraction, the concentra-

http://www.elsevier.com/locate/trac 159
Trends Trends in Analytical Chemistry, Vol. 46, 2013

tions determined in this study probably underestimate Appendix A. Supplementary data

the real occurrence of PCDDs/PCDFs in air. Hence, the
concentrations in Africa and Latin America are probably Supplementary data associated with this article can be
even higher than in Europe. Combustion sources are found, in the online version, at http://dx.doi.org/
possibly the reason for the emissions of PCDDs/PCDFs. In 10.1016/j.trac.2012.05.011.
Africa and Latin America, open waste burning, partic-
ularly e-waste burning, may represent relevant sources
of PCDDs/PCDFs [45]. Combustion of biomass, forest References
fuel, and sugarcane can also lead to the formation of [1] H. Fiedler, Persistent Organic Pollutants, Springer Heidelberg,
PCDDs/PCDFs, though, to a lesser extent than burning Dordrecht, London, New York, USA, 2003.
[2] United Nations Environmental Programme, 2009 (http://
synthetic materials [46,47]. Additionally, open dumping chm.pops.int/Convention/ConventionText/tabid/2232/Default.
sites in developing countries, particularly open landfills aspx#LiveContent).
of municipal solid waste, were reported to represent [3] S. Corsolini, J. Chromatogr., A 1216 (2009) 598.
relevant sources of PCDDs/PCDFs [48]. [4] H. Hung, R. Kallenborn, K. Breivik, Y.S. Su, E. Brorstrom-Lunden,
K. Olafsdottir, J.M. Thorlacius, S. Leppanen, R. Bossi, H. Skov, S.
Mano, G.W. Patton, G. Stern, E. Sverko, P. Fellin, Sci. Total
5. Conclusions Environ. 408 (2010) 2854.
[5] R.W. MacDonald, L.A. Barrie, T.F. Bidleman, M.L. Diamond, D.J.
Gregor, R.G. Semkin, W.M.J. Strachan, Y.F. Li, F. Wania, M.
This study has confirmed that PASs using PUF disks Alaee, L.B. Alexeeva, S.M. Backus, R. Bailey, J.M. Bewers, C.
represent a suitable sampling technique for the deter- Gobeil, C.J. Halsall, T. Harner, J.T. Hoff, L.M.M. Jantunen, W.L.
mination of POPs in the atmosphere, particularly in Lockhart, D. Mackay, D.C.G. Muir, J. Pudykiewicz, K.J. Reimer,
developing countries and countries with economy in J.N. Smith, G.A. Stern, W.H. Schroeder, R. Wagemann, M.B.
Yunker, Sci. Total Environ. 254 (2000) 93.
transition, due to their low cost, easy handling, and
[6] D. Muir, B. Braune, B. DeMarch, R. Norstrom, R. Wagemann, L.
independence of power supply. This study showed that Lockhart, B. Hargrave, D. Bright, R. Addison, J. Payne, K. Reimer,
continuous monitoring of POPs in background sites of Sci. Total Environ. 230 (1999) 83.
non-industrialized countries is important to provide a [7] K. Weber, H. Goerke, Chemosphere 53 (2003) 667.
basis for evaluating the effectiveness of the Stockholm [8] S. Du, S.J. Wall, D. Cacia, L.A. Rodenburg, Environ. Sci. Technol.
43 (2009) 1287.
Convention on POPs.
[9] R. Gioia, S. Eckhardt, K. Breivik, F.M. Jaward, A. Prieto, L.
The issue of waste export from industrialized countries Nizzetto, K.C. Jones, Environ. Sci. Technol. 45 (2011) 1349.
to Africa and Asia deserves more attention and action. [10] R. Gioia, L. Nizzetto, R. Lohmann, J. Dachs, C. Temme, K.C. Jones,
Also, more extensive knowledge transfer to developing Environ. Sci. Technol. 42 (2008) 1416.
countries and countries with economy in transition [11] A.K. Halse, M. Schlabach, S. Eckhardt, A. Sweetman, K.C. Jones,
K. Breivik, Atmos. Chem. Phys. 11 (2011) 1549.
about adequate combustion techniques and handling
[12] F.M. Jaward, N.J. Farrar, T. Harner, A.J. Sweetman, K.C. Jones,
practice of hazardous waste is needed to reduce their Environ. Sci. Technol. 38 (2004) 34.
atmospheric emissions of PCDDs/PCDFs. Regarding DDT, [13] T.M. Jaward, G. Zhang, J.J. Nam, A.J. Sweetman, J.P. Obbard, Y.
the high concentrations observed in ambient air in some Kobara, K.C. Jones, Environ. Sci. Technol. 39 (2005) 8638.
regions currently using DDT call for further support for [14] J. Kla´nova´, P. Čupr, I. Holoubek, J. Boru˚ vkova´, P. Přibylova´, R.
Karesˇ, J. Kohoutek, A. Dvorska´, T. Tomsˇej, T. Ocelka, RECETOX
sustainable alternatives to DDT for malaria control.
MU, Czech Republic, Brno, 2008 (ISBN 9788021047396).
Finally, knowledge transfer, and training on methods [15] K. Pozo, T. Harner, S.C. Lee, R.K. Sinha, B. Sengupta, M. Loewen,
to analyze POPs in environmental samples would enable V. Geethalakshmi, K. Kannan, V. Volpi, Environ. Pollut. 159
developing countries to perform their own monitoring (2011) 646.
studies, and to reduce and to prevent environmental [16] K. Pozo, T. Harner, S.C. Lee, F. Wania, D.C.G. Muir, K.C. Jones,
Environ. Sci. Technol. 43 (2009) 796.
pollution by POPs more effectively.
[17] K. Pozo, T. Harner, M. Shoeib, R. Urrutia, R. Barra, O. Parra, S.
Focardi, Environ. Sci. Technol. 38 (2004) 6529.
[18] E. Abad, J. Caixach, J. Rivera, L. Gustems, G. Massague, O. Puig,
Acknowledgements Sci. Total Environ. 334 (2004) 279.
Jacob de Boer and Heather Leslie (IVM, Free University [19] I. Holoubek, J. Klanova, J. Jarkovsky, J. Kohoutek, J. Environ.
Amsterdam, the Netherlands), Gene Martrat (IDÆA- Monit. 9 (2007) 557.
CSIC, Barcelona, Spain), and Yingming Li (RCEES, Chi- [20] A. Katsoyiannis, R. Gioia, A.J. Sweetman, K.C. Jones, Environ. Sci.
Technol. 44 (2010) 5735.
nese Academy of Sciences, Beijing, P.R. China) are
[21] E. Menichini, N. Lacovella, F. Monfredini, L. Turrio-Baldassarri,
acknowledged for their support for the sample analysis. Chemosphere 69 (2007) 422.
The generation of data has been made possible through [22] G. Umlauf, E.H. Christoph, S.J. Eisenreich, G. Mariani, B. Paradiz,
UNEP implemented global monitoring projects founded I. Vives, Environ. Sci. Pollut. Res. Int. 17 (2010) 462.
by GEF (Global Environmental Facility) and SAICM-QSP [23] United Nations Environment Programme (http://www.chem.u-
nep.ch/Pops/GMP/Report%20UNEP%20PAS%20Air%20Moni-
(Strategic Approach to International Chemicals Man-
toring%202010-2011.pdf.)
agement - Quick Start Programme). [24] M. Shoeib, T. Harner, Environ. Sci. Technol. 36 (2002) 4142.

160 http://www.elsevier.com/locate/trac
Trends in Analytical Chemistry, Vol. 46, 2013 Trends

[25] C. Chaemfa, E. Wild, B. Davison, J.L. Barber, K.C. Jones, J. Environ. [38] F. Demaria, Ecolog. Econom. 70 (2010) 250.
Monit. 11 (2009) 1135. [39] M.S. Hossain, S.R. Chowdhury, S.M.A. Jabbar, S.M. Saifullah,
[26] L. Tuduri, T. Harner, H. Hung, Environ. Pollut. 144 (2006) 377. M.A. Rahman, Chiang Mai J. Sci. 35 (2008) 370.
[27] L. Melymuk, M. Robson, P.A. Helm, M.L. Diamond, Atmos. [40] G. Zhang, P. Chakraborty, J. Li, P. Sampathkumar, T. Balasubr-
Environ. 45 (2011) 1867. amanian, K. Kathiresan, S. Takahashi, A. Subramanian, S.
[28] D. Mackay, Handbook of Physical-Chemical Properties and Tanabe, K.C. Jones, Environ. Sci. Technol. 42 (2008) 8218.
Environmental Fate for Organic Chemicals, Second Edition., CRC [41] R. Weber, C. Gaus, M. Tysklind, P. Johnston, M. Forter, H. Hollert,
Press/Taylor & Francis, Boca Raton, FL, USA, 2006. E. Heinisch, I. Holoubek, M. Lloyd-Smith, S. Masunaga, P.
[29] M. van den Berg, L. Birnbaum, A.T.C. Bosveld, B. Brunstrom, P. Moccarelli, D. Santillo, N. Seike, R. Symons, J.P.M. Torres, M.
Cook, M. Feeley, J.P. Giesy, A. Hanberg, R. Hasegawa, S.W. Verta, G. Varbelow, J. Vijgen, A. Watson, P. Costner, J. Woelz, P.
Kennedy, T. Kubiak, J.C. Larsen, F.X.R. van Leeuwen, A.K.D. Wycisk, M. Zennegg, Environ. Sci. Pollut. Res. Int. 15 (2008) 363.
Liem, C. Nolt, R.E. Peterson, L. Poellinger, S. Safe, D. Schrenk, D. [42] K. Breivik, R. Gioia, P. Chakraborty, G. Zhang, K.C. Jones,
Tillitt, M. Tysklind, M. Younes, F. Waern, T. Zacharewski, Environ. Sci. Technol. 45 (2011) 9154.
Environ. Health Perspect. 106 (1998) 775. [43] United Nations Environmental Programme, Expert Group to
[30] J.V. de Assuncao, U.R. Pesquero, R.E. Bruns, L.R.F. Carvalho, assess DDT production and use and its alternatives – First
Chemosphere 58 (2005) 1391. meeting, 2007 (http://chm.pops.int/Portals/0/Repository/
[31] B. Cardenas, J. Ferrario, C. Byrne, R. Basaldud, S. Benitez, H. DDTEG1/UNEP-POPS-DDT-EG.1-5.English.PDF). (http://chm.pop-
Wöhrnschimmel, A. Ortinez, F. Angeles, I. Galindo, G. Gold, E. s.int/Portals/0/Repository/DDTEG1/UNEP-POPS-DDT-EG.1-5.
Herrera, V. Macı́as, C. Marquez, T. Ortun˜ o, A. Tellez, C. Welsh, English.PDF).
Organohalogen Compd. 73 (2011) 1030. [44] M. Over, B. BakoteÕe, R. Velayudhan, P. Wilikai, P.M. Graves, Am.
[32] B.H. Aristizabal, C.M. Gonzalez, L. Morales, M. Abalos, E. Abad, J. Trop. Med. Hyg. 71 (2004) 214.
Chemosphere 85 (2011) 170. [45] T. Zhang, H. Fiedler, G. Yu, G. Solorzano Ochoa, W.F. Carroll Jr.,
[33] K. Breivik, A. Sweetman, J.M. Pacyna, K.C. Jones, Sci. Total B.K. Gullett, S. Marklund, A. Touati, Chemosphere 84 (2011)
Environ. 377 (2007) 296. 994.
[34] B.H. Robinson, Sci. Total Environ. 408 (2009) 183. [46] R.R. Black, C.P. Meyer, A. Touati, B.K. Gullett, H. Fiedler, J.F.
[35] H.T. Shen, G.Q. Ding, G.G. Han, X.F. Wang, X.M. Xu, J.L. Han, Mueller, Chemosphere 83 (2011) 1331.
X.M. Lou, C.J. Xu, D.L. Cai, Y.H. Song, W. Lu, Chemosphere 80 [47] R.R. Black, C.P. Meyer, A. Touati, B.K. Gullett, H. Fiedler, J.F.
(2010) 170. Mueller, Environ. Int. 38 (2012) 62.
[36] J.Q. Zhang, Y.S. Jiang, J. Zhou, B. Wu, Y. Liang, Z.Q. Peng, D.K. [48] N.H. Minh, T.B. Minh, M. Watanabe, T. Kunisue, I. Monirith, S.
Fang, B. Liu, H.Y. Huang, C. He, C.L. Wang, F.N. Lu, Environ. Sci. Tanabe, S. Sakai, A. Subramanian, K. Sasikumar, P.H. Viet, B.C.
Technol. 44 (2010) 3956. Tuyen, T.S. Tana, M.S. Prudente, Environ. Sci. Technol. 37
[37] K.A. Asante, S. Adu-Kumi, K. Nakahiro, S. Takahashi, T. Isobe, A. (2003) 1493.
Sudaryanto, G. Devanathan, E. Clarke, O.D. Ansa-Asare, S.
Dapaah-Siakwan, S. Tanabe, Environ. Int. 37 (2011) 921.

http://www.elsevier.com/locate/trac 161