Semipermeable membrane devices as
integrative tools for monitoring nonpolar
aromatic compounds in air
by
Hanna Söderström
Akademisk avhandling
Som med tillstånd av rektorsämbetet vid Umeå universitet för erhållande av
Filosofie Doktorsexamen vid Teknisk-naturvetenskapliga fakulteten i Umeå,
framlägges till offentlig granskning vid Kemiska Institutionen, hörsal KB3B1 i
KBC, fredagen den 10 december, 2004, kl. 10.00.
Fakultetsopponent: Associate Professor Jochen F. Müller, National Research
Centre of Environmental Toxicology (EnTox), University of Queensland,
Australia.
Copyright © 2004 Hanna Söderström
Cover illustration by Leif Lindgren, Älvsbyn
Environmental Chemistry
Department of Chemistry
Umeå University
SE-901 87 Umeå
Sweden
ISBN 91-7305-782-7
Printed in Sweden by VMC, KBC, Umeå University, Umeå 2004.
Semipermeable membrane devices as integrative tools for monitoring nonpolar
aromatic compounds in air
Hanna Söderström, Environmental Chemistry, Umeå University, Umeå, Sweden.
Abstract: Air pollutants pose a high risk for humans, and the environment, and this pollution
is one of the major environmental problems facing modern society. Active air sampling is the
technique that has been traditionally used to monitor nonpolar aromatic air pollutants.
However, active high volume samplers (HiVols) require a power supply, maintenance and
specialist operators, and the equipment is often expensive. Thus, there is a need to develop
new, less complicated sampling techniques that can increase the monitoring frequency, the
geographical distribution of the measurements, and the number of sites used in air
monitoring programs. In the work underlying this thesis, the use of semipermeable
membrane devices (SPMDs) as tools for monitoring gas phase concentrations of nonpolar
aromatic compound was evaluated using the compound classes polychlorinated biphenyls
(PCBs), polycyclic aromatic hydrocarbons (PAHs), alkylated PAHs (alkyl-PAHs) and nitrated
PAHs (nitro-PAHs) as test compounds.
High wind-speeds increased the uptake and release in SPMDs of PAHs and PCBs with log
KOA values > 7.9, demonstrating that the uptake of most nonpolar aromatic compounds is
controlled by the boundary layer at the membrane-air interface. The use of a metal umbrella
to shelter the SPMDs decreased the uptake of PAHs and PCBs by 38 and 55 percent,
respectively, at high wind/turbulence, and thus reduced the wind effect. Further, the use of
performance reference compounds (PRCs) to assess the site effect of wind on the uptake in
SPMDs reduced the between-site differences to less than 50 percent from as much as three
times differences in uptake of PCBs and PAHs. However, analytical interferences reduced the
precision of some PRCs, showing the importance of using robust analytical quality control.
SPMDs were shown to be efficient samplers of gas phase nonpolar aromatic compounds,
and were able to determine local, continental and indoor spatial distributions of PAHs, alkylPAHs and nitro-PAHs. In addition, the use of the SPMDs, which do not require electricity,
made sampling possible at remote/rural areas where the infrastructure was limited. SPMDs
were also used to determine the source of PAH pollution, and different approaches were
discussed. Finally, SPMDs were used to estimate the importance of the gas phase exposure
route to the uptake of PAHs in plants. The results demonstrate that SPMDs have several
advantages compared with HiVols, including integrative capacity over long times, reduced
costs, and no need of special operators, maintenance or power supply for sampling.
However, calibration data of SPMDs in air are limited, and spatial differences are often only
semi-quantitatively determined by comparing amounts and profiles in the SPMDs, which
have limited their use in air monitoring programs. In future work, it is therefore important
that SPMDs are properly sheltered, PRCs are used in the sampling protocols, and that
calibrated sampling rate data, or the SPMD-air partition data, of specific compounds are
further developed to make determination of time weighted average (TWA) concentrations
possible.
Keywords: air pollution, alkyl-PAHs, atmosphere, bioavailability, boundary layer, diffusive
sampling, emissions, integrative, membrane, monitoring, nitro-PAHs, PAHs, particles,
passive samplers, PCBs, plants, PRCs, release rate, sampler design, sources, SPMDs, traffic,
uptake rate, wind effect, wood burning.
ISBN 91-7305-782-7
Semipermeabla membraner som integrerande mätverktyg vid miljöövervakning
av opolära aromatiska föreningar i luft
Hanna Söderström, Miljökemi, Umeå Universitet, Umeå, Sverige.
Sammanfattning: Luftföroreningar utgör en stor risk för människa och miljö, och är ett av
de största miljöproblemen i vårt samhälle. Traditionellt har aktiv luftprovtagning används vid
miljöövervakning av opolära aromatiska luftföroreningar. Aktiv högvolymprovtagning kräver
dock elektricitet, underhåll och operatörer med specialistkompetens, och utrustningen är ofta
dyr. Det finns därför ett behov av nya, mindre komplicerade provtagningstekniker som kan
öka mätfrekvensen, den geografiska spridningen samt antalet provtagningsplatser som
används i miljöövervakningsprogram. I det arbete som presenteras i denna avhandling
undersöktes användningen av semipermeabla membraner (SPMDer) som mätverktyg av
opolära aromatiska föreningar i gasfas. Ämnesklasserna polyklorerade bifenyler (PCBer),
polycykliska aromatiska kolväten (PAHer), alkylerade PAHer (alkyl-PAHer) och nitrerade
PAHer (PAHer) användes som testsubstanser.
Höga vindhastigheter ökade upptaget samt avgången av PAHer och PCBer med log KOA
värden > 7.9, vilket visar att upptaget i SPMDer av de flesta opolära aromatiska föreningar
kontrolleras av diffusionslagret som bildas vid gränsytan mellan SPMD och luft. Då ett
metallparaply användes för att skydda de SPMDerna minskades PAH- och PCB-upptaget
med 38 samt 55 procent, vilket visar att metallparaplyet reducerade effekten av vind.
Användningen av sk performance reference compounds (PRCs) för att bedömma effekten av
vind på upptaget i SPMDer reducerade variationen mellan SPMDer som utsatts för olika
vindhastigheter till mindre än 50 procent från upp till tre gångers skillnad i PAH- och PCBupptaget. Analytiska interferenser reducerade dock precisionen hos några PRCs vilket visar
vikten av att använda robust analytisk kvalitetskontroll.
SPMDer visade sig vara effektiva provtagare av opolära aromatiska föreningar i gasfas och
kunde användas för att bestämma PAHers, alkyl-PAHers samt nitro-PAHers spridning i luft,
lokal, kontinental samt inomhus. Eftersom SPMDer ej kräver elektricitet möjliggjordes även
provtagning på otillgängliga plaster där infrastrukturen var begränsad. SPMDer kunde även
användas för att bestämma källor till PAH-förorening i luft, och olika tillvägagångssätt för
detta ändamål diskuterades. Slutligen användes SPMDer för att bedömma betydelsen av
gasfasexponeringen för PAH-upptaget i växter. Dessa resultat visar att SPMDer har många
fördelar i jämförelse med högvolymprovtagare som integrerande kapacitet under långa
tidsperioder, reducerade kostnader, och ej krav på operatörer med specialistkompetens,
elektricitet och underhåll vid provtagning. Tillgängliga kalibreringsdata för SPMDer i luft är
dock begränsande, och haltskillnader bestäms oftast endast semikvantitativt genom att
jämföra halter och profiler funna i SPMDer, vilket har begränsat deras användning i
miljöövervakningsprogram. I framtida arbeten är det därför viktigt att de SPMDerna är
ordentligt skyddade, att PRCs används i provtagningsprotokollen, och att fler kalibrerade
provtagningshastigheter, eller SPMD-luftfördelningsdata för specifika föreningar tas fram så
att tidsvägsmedelkoncentrationer kan bestämmas.
Nyckelord: alkyl-PAHer, atmosfär, avgångshastighet, biotillgänglighet, diffusionslager,
diffusionsprovtagning, emissioner, integrerande, källor, luftförorening, membran,
miljöövervakning, nitro-PAHer, PAHer, partiklar, passiva provtagare, PCBer, PRCs,
provtagardesign, SPMDer, trafik, upptagshastigheter, vindeffekt, vedeldning, växter.
ISBN 91-7305-782-7
Det krävs ett helt nytt sätt
att tänka för att lösa de problem
vi skapat med det gamla
sättet att tänka.
Today´s problems cannot
be solved with the same thinking
that created them.
Albert Einstein
i
ii
LIST OF PAPERS
This thesis is based on the following Papers, which are referred to in the text
by their roman numbers as listed below.
I. Söderström H.S., and Bergqvist P-A., 2003. Polycyclic aromatic
hydrocarbons in a semiaquatic plant and semipermeable membrane
devices exposed to air in Thailand. Environmental Science and Technology.
37, 47-52.
II. Söderström H.S., and Bergqvist P-A. Wind effects on passive air
sampling of PAHs and PCBs. Accepted for publication in Bulletin of
Environmental Contamination and Toxicology.
III. Söderström H.S., and Bergqvist P-A., 2004. Passive air sampling using
semipermeable membrane devices at different wind-speeds in situ
calibrated by performance reference compounds. Environmental Science
and Technology. 38, 4828-4834.
IV. Söderström H.S., Hajšlová J., Siegmund B., Kocan A., Obiedzinski
M.W., Tysklind M., and Bergqvist P-A. PAHs and nitrated PAHs in
air of five European countries determined using SPMDs as passive
samplers. Accepted for publication in Atmospheric Environment.
V. Strandberg B., Söderström H.S, Gustafson P., Sällsten G., Bergqvist
P-A., and Barregård L., 2004. Semipermeable membrane devices as
passive samplers to determine polycyclic hydrocarbons (PAHs) in
indoor environments. Organohalogen Compounds. 66, 44-49.
All papers are reproduced with the kind permission of the publishers of the
respective journals.
iii
iv
ABBREVIATIONS
alkyl-PAH
AU
13
C
CCC
CERC
CO
CV
CZ
d
DDT
dw
EAFs
EMEP
EPA
FC
GC/MS
GPC
H
2
H
HCB
HCH
HiVols
HR
KAW
ke
KOA
KOW
KSA
KSPMD
LB
LDPE
LLE
LR
LRTAP
MeVols
nitro-PAH
1-NN
2-NN
Alkylated PAH
Austria
Carbon 13-labelled compound
Chemical coordinating center
Columbia environmental research center
Carbon monoxide
Coefficient of variance
Czech Republic
Day (24 hours)
Dichlorodiphenyltrichloroethane
Dry weight
Exposure adjustment factors
European monitoring and evaluation program
Environmental protection agency
Field control
Gas chromatography/mass spectrometry
Gel permeation chromatography
Henry’s law constant
Deuterated compound
Hexachlorobenzene
Hexachlorocyclohexane
Active high volume sampler
High-resolution
Air-water partition coefficient
Exchange rate constant
Octanol-air partition coefficient
Octanol-water partition coefficient
SPMD-air partition coefficient
SPMD equilibrium partition coefficient (also, SPMD-water
partition coefficient)
Laboratory blank
Low-density polyethylene
Liquid-liquid extraction
Low-resolution
Long range transboundary air pollution
Active medium volume sampler
Nitrated PAH
1-nitronaphthalene
2-nitronaphthalene
v
NO2
OC
PAH
PAS
PBDE
PCA
PCB
PCDD
PCDF
PM
POGs
POP
PRC
PUF
QA
QC
RS
SE
SK
SO2
SPE
S-PL
SPMD
SPME
SVOC
TWA
USGS
VOC
ww
W-PL
Nitrogen dioxide
Organochlorine pesticide
Polycyclic aromatic hydrocarbon
Passive air samplers
Polybrominated diphenyl ether
Principal component analysis
Polychlorinated biphenyl
Polychlorinated dioxin
Polychlorinated furan
Particulate matter
Polymer-coated glass sampler
Persistent organic pollutant
Performance reference compound
Polyurethane foam
Quality assurance
Quality control
Sampling rate
Sweden
Slovakia
Sulphur dioxide
Solid phase extraction
Poland summer 2000
Semipermeable membrane device
Solid phase microextraction
Semi volatile organic compound
Time weighted average
United States geological survey’s
Volatile organic compound
Wet weight
Poland winter 1999
vi
TABLE OF CONTENTS
1
1. INTRODUCTION
3
1.1 Aims and scope
5
2. PASSIVE AIR SAMPLING
7
2.1 Historical use of PAS as monitoring tools
8
2.1.1 Passive air sampling of nonpolar aromatic compounds
9
2.2 Passive sampling using semipermeable membrane devices
2.2.1 Sampling principles and factors influencing the sampling rate 12
15
2.2.2 Target compounds
16
2.2.3 Fields of application
2.2.4 Theory and mathematical models of SPMD sampling
18
21
2.2.4.1 The PRC model
3. COMPOUNDS SAMPLED AND THEIR PHYSICAL/CHEMICAL
23
PROPERTIES
23
3.1 Polycyclic aromatic hydrocarbons (PAHs)
24
3.2 Polychlorinated biphenyls (PCBs)
27
4. SAMPLING AND CHEMICAL ANALYSIS
27
4.1 Design of the sampler
28
4.2 Sampling procedure and approach
28
4.2.1 The PRC approach
29
4.2.2 Cleaning of the sampling equipment
29
4.3 Extraction, cleanup and analysis
30
4.4 Quality assurance and quality control (QA/QC)
32
4.5 Data interpretation
32
4.5.1 Semi-quantitative and quantitative approach
4.5.2 Multivariate data analysis using principal component analysis
33
(PCA)
5. FACTORS INFLUENCING THE UPTAKE OF COMPOUNDS IN
35
SPMDs FROM AIR
36
5.1 Physical/chemical properties of the compounds sampled
37
5.1.1 RS of PCBs in air
38
5.2 Environmental variables
39
5.2.1 Wind-speed/turbulence
40
5.2.1.1 The wind effect test method
5.2.1.2 Wind effect at high wind-speeds
42
44
5.2.1.3 Wind effect in field
44
5.2.2 Temperature
45
5.2.3 Particles
47
5.2.4 UV sunlight
6. APPROACHES TO REDUCE THE SITE EFFECTS OF
49
ENVIRONMENTAL VARIABLES
49
6.1 The use of shelter
50
6.1.1 Reduction of the wind effect by shelter
6.2 The use of performance reference compounds (PRCs)
52
52
6.2.1 PRCs´ utility for assessing wind effect
57
6.2.2 PRCs´ utility for assessing site effects in air
59
7. FIELDS OF APPLICATION AND SAMPLING STATEGIES
59
7.1 Determination of local/regional atmospheric distribution
61
7.2 Determination of continental atmospheric distribution
65
7.3 Determination of pollution sources
65
7.3.1 The use of individual compound ratios
7.3.1.1 Considerations involved with the use of the individual compound
67
ratios
68
7.3.2 The use of total PAH-patterns
7.3.3 Determination of residential wood burning as a source of indoor
69
PAH exposure
7.4 Estimation of the consequences of gas phase PAH exposure for
72
plants
8. CONCLUSIONS AND FUTURE RESEARCH POSSIBILITIES 77
79
9. ACKNOWLEDGEMENTS
81
REFERENCES
1. Introduction
1. INTRODUCTION
Air pollutants pose a high risk for humans, and the environment, and are one
of the major environmental problems facing modern society. For instance,
relationships between exposure to outdoor air pollutants and daily mortality,
morbidity, hospital admissions, and lung function changes have been found in
many studies (e.g., Forsberg et al., 1998; Aga et al., 2003; Sunyer et al., 2003;
Zanobetti et al., 2003; Chen et al., 2004). Anthropogenic air pollution has
occurred since ancient times, but increased enormously in the 20th century
following the industrial revolution in Europe, and continued to increase
significantly until the 1970s, due to increases in population, industrialization,
urbanization and traffic (Figure 1). Today, many countries have reduced their
largest local point sources, such as emissions from industry, while the
diffusive sources such as traffic continue to increase as a result of our modern
living habits. Air pollution has also become a global problem since many of
the air pollutants are transported over long ranges with atmospheric and
ocean currents. Thus, even if a country reduces its major emissions, air
pollution can continue to increase. In order to control the air quality, and thus
the human exposure through air, air monitoring is highly important.
Figure 1. Illustration of the continuous pollution of the atmosphere due to our
modern living habits which has become one of the most serious environmental
problems facing today’s society.
Air monitoring on a national level is often regulated by international
conventions. In Europe, one of the major monitoring programs is the
European Monitoring and Evaluation Program (EMEP). This monitoring
program was established after the convention on Long Range Transboundary
Air Pollution (LRTAP) was signed in 1979 (Geneva, November 1979). The
main objective of EMEP is to regularly provide Governments and subsidiary
1
1. Introduction
bodies under the LRTAP convention with useful research results for the
development and evaluation of international protocols on the reduction of
atmospheric emissions that are discussed and detailed within the convention.
The EMEP has three main purposes; to collect emission data, to measure air
and precipitation quality, and to model atmospheric transport and deposition.
The focus are on acidifying and eutrophying compounds, ground level ozone,
heavy metals, persistent organic pollutants (POPs) or semi volatile organic
compounds (SVOCs) (classed in this thesis as nonpolar aromatic
compounds), volatile organic compounds (VOCs) and particulate matter
(PM). The Chemical Coordinating Centre (CCC) co-ordinates the EMEP
measurements of chemical air quality and precipitation. In 2002, six countries
measured POPs in air and aerosols in a total of eight sites (Aas and Breivik,
2004). Both the distribution of the sites across Europe and the number of
sites at which POPs were measured in this monitoring program were
considered by the CCC as insufficient (Aas and Breivik, 2004). Thus, the real
spatial variability in the atmospheric concentrations of POPs across Europe
remains rather unknown.
The standard methods for air sampling of nonpolar aromatic compounds like
pesticides, polychlorinated biphenyls (PCBs) and polycyclic aromatic
hydrocarbons (PAHs) recommended by CCC/EMEP are active
high/medium volume samplers (HiVols/MeVols) including a glass fibre filter
(which collects particle-bound compounds) followed by two polyurethane
foam (PUF) plugs or XAD-2 (which retain gaseous compounds) coupled to a
pump (Anonymous, 2001). Active air sampling is the technique that has been
traditionally used in air monitoring. The disadvantages with active volume
samplers are that they require power and maintenance, the sampling
equipment is often both expensive and complicated, and they have to be
operated by specially trained personnel. The high cost of the sampling
equipment, the power demands and the requirement for maintenance by
specialist operators restrict the number and location of the sites that can be
monitored. Furthermore, active sampling is seldom integrative for more than
48 h (see, for instance, reported sampling frequencies by Aas and Breivik
(2004)), which makes it difficult to measure episodic pollutant releases and
time weighted average (TWA) concentrations. Finally, in active sampling the
total amounts of air pollutants (of both the gas phase and the particleassociated compounds) are collected although one or the other of these
phases may pose little or no primary risk to living organisms and humans.
Thus, there is a need to develop new, less complicated sampling techniques
that can increase the monitoring frequency, the geographical distributions of
2
1. Introduction
the measurements, and the number of sites used in air monitoring programs
of POPs.
1.1 Aims and scope
In the work described in this thesis, the use of semipermeable membrane
devices (SPMDs) as tools for monitoring the gas phase concentrations of
nonpolar aromatic compounds in air was evaluated. The work focused on the
nonpolar aromatic compounds PCBs, PAHs, alkylated PAHs (alkyl-PAHs)
and nitrated PAHs (nitro-PAHs). The main objective was to further develop
the SPMDs as passive air samplers (PAS) to provide an alternative sampling
method to the conventionally used active volume samplers. Briefly, this thesis
discusses the uptake and release rates of the gas phase of nonpolar aromatic
compounds in SPMDs, the factors influencing the SPMD sampling rate and
the performance of SPMDs in different fields of application.
More specifically, the work described in Papers II and III tested the
hypothesis that high wind-speeds/turbulence increase the uptake in SPMDs.
Further, the use of performance reference compounds (PRCs) to calibrate the
SPMD sampling rate in situ and thus assess site effects at different windspeeds/turbulence was tested (Papers II and III). Paper III also tested the
hypothesis that shelters protecting the SPMDs reduced the wind/turbulence
exposure, and thus decreased the uptake and release rates of target analytes.
In unpublished work, the RS values of a number of PCBs for SPMDs in air
were estimated by comparing SPMD samplers with conventionally used
MeVols. Finally, in Papers I, IV and V the scope for using SPMDs to
determine the local, continental and indoor spatial distribution, respectively,
of PAHs and some of their substituted derivates, was examined. The work
described in Papers I, IV and V also aimed to determine potential point
sources of PAHs like traffic and residential heating systems, and to investigate
different strategies to detect the specific emission sources. The consequence
of gas phase exposure, sampled by SPMDs, on the uptake of PAHs in a semiaquatic plant were also investigated (Paper I).
3
4
2. Passive air sampling
2. PASSIVE AIR SAMPLING
Passive air sampling is a sampling method that has advantages in spatial
surveys because of its low expense, simplicity, long-term operation and the
fact that it does not require electricity. The general sampling principles of
PAS and their historical use as monitoring tools are outlined below. In
addition, the use of SPMDs as passive samplers is discussed in detail.
The definition of a passive or diffusive sampler is a “device which is capable
of taking samples of gases or vapours from the atmosphere at a rate
controlled by a physical process such as gaseous diffusion through a static air
layer or a porous material and/or permeation through a membrane, but
which does not involve the active movement of air through the device”
(Anonymous, 2002a). Active sampling of air is normally achieved by pumps.
In European standards, the mass of an analyte taken up by a diffusive sampler
is described by (Anonymous, 2003)
mS=A×D×(ρ1-ρ2)×t
l
Eq. 1
where mS is the mass of an analyte which is sorbed by diffusion in pg, A is the
cross-sectional area of the diffusion path or the equivalent sorption surface in
cm2, D is the diffusion coefficient of the analyte in cm2 ⋅ min-1, ρ1 is the
concentration of the analyte at the beginning of the diffusive layer in µg ⋅ m-3,
ρ2 is the concentration of the analyte at the end of the diffusive layer in µg ⋅
m-3, t is the exposure time in min, and l is the length of the static air layer in
the sampler or equivalent for permeation types, in cm (Figure 2).
Ideally, ρ1 is equal to the air concentration of the analyte outside the diffusive
sampler and ρ2 is reduced to zero due to sorption or chemical reactions in the
sorbent. A ρ1 close to or equal to zero (“zero sink”-condition) serves as the
driving force of the diffusion across l.
5
2. Passive air sampling
A
l
ρ
ρ2
B
ρ1
Figure 2. Schematic diagram of diffusive sampling and parameters influencing the
process (modified from Anonymous, 2003).
The diffusive uptake rate, U, in cm2 ⋅ min-1, is obtained from knowledge of
either the physical parameters of the diffusion barrier (Anonymous 2002b)
according to
U = ms =A×D
ρ×t
l
Eq. 2
or by comparison with another analyte for which the diffusive uptake rate, U2,
and the diffusion coefficient of the analyte, D2, is known (Anonymous,
2002b) according to
U1=D1×U2
D2
Eq. 3
The major advantages of PAS over active samplers are their low expense and
simplicity, since neither a power supply nor specialist operators are required
for routine monitoring, and they can operate for long time periods. Thus,
PAS can facilitate sampling campaigns at multiple sites simultaneously on
local, regional or global levels. They may also be the only option when
sampling is required in remote locations, such as natural reserves and
mountain areas where there is no electricity supply. The mass uptake of many
PAS is linear and integrative during the entire exposure period and TWA
concentrations over long time periods can be assessed. Integrative diffusive
sampling has several advantages. In the integrative sampling approach
episodic pollutant releases are detected to a higher degree than during active
grab sampling. TWA concentrations are also more relevant in exposure
assessments because they reflect the total exposure during a given time
period. Passive sampling is also preferable to active low volume sampling
6
2. Passive air sampling
when the ambient air concentrations are low because integrative sampling
concentrates compounds present in trace or ultra-trace levels to levels above
detection limits. However, POPs bound to ultra-fine particles (diameter < 1
µm), which can have toxic effects on the respiratory organs of humans and
wildlife, are not measured with PAS. In addition, the sampling rate of PAS is
affected by the sampling conditions, especially factors like windspeed/turbulence and temperature, and it is important to ensure that the
effect of these parameters is reduced as much as possible and calibrated.
2.1 Historical use of PAS as monitoring tools
Passive sampling was employed for the first time in 1927 (Gorecki and
Namiesnik, 2002). The cited passive sampler was used to semi-quantitatively
determine the atmospheric concentration of carbon monoxide (CO) (Gorecki
and Namiesnik, 2002). However, there was a delay of almost 50 years (until
1973) before diffusive sampling was described in mathematical terms,
sampling rates were calibrated, and two types of passive samplers could be
used to quantitatively determine the air concentration of nitrogen dioxide
(NO2) and sulphur dioxide (SO2) (Gorecki and Namiesnik, 2002).
Since then, these and new types of PAS have been further developed (e.g.,
Berglund et al., 1992; Cao and Hewitt, 1993; Krochmal and Kalina, 1997;
Wideqvist et al., 2003). To date, PAS have been routinely employed in
occupational exposure assessments and in ambient, outdoor air monitoring of
a variety of volatile organic compounds (VOCs) like NO2, SO2, ozone, and
benzene (see, for example, Sällsten et al., 2001; Stevenson et al., 2001;
Bytnerowicz et al., 2002; Modig et al., 2002; Kruså et al., 2004). For instance,
the Swedish environmental protection agency has assessed the ambient
human exposure of NO2, benzene, butadiene, formaldehyde and acetaldehyde
by using personal passive sampling (an approach in which the sampler is
carried by people judged to be at risk of exposure) (Sällsten et al., 2001;
Modig et al., 2002; Kruså et al., 2004). In total, 40-60 persons were randomly
selected from the populations of each of three cities Gothenburg (2000),
Umeå (2001) and Stockholm (2002 and 2003), and integrative personal
passive sampling was performed for a week in each case (Sällsten et al., 2001;
Modig et al., 2002; Kruså et al., 2004).
7
2. Passive air sampling
2.1.1 Passive air sampling of nonpolar aromatic compounds
The local and continental atmospheric distribution of nonpolar aromatic
compounds was first monitored by collecting different plant materials such as
tree bark (e.g., Simonich and Hites, 1995), leaves and grass (e.g., Müller et al.,
2001), pine needles (e.g., Tremolada et al., 1996; Kylin and Sjödin, 2003;
Hellström et al., 2004; Lehndorff and Schwark, 2004) and liches (e.g.,
Ockenden et al., 1998b). In biomonitoring-based air sampling, the spatial
pollution distribution in the plant tissues and the atmosphere are assumed to
be closely related. Kylin and Sjödin (2003), for instance, analysed pine needle
wax collected at various sites from the central to northern parts of Europe, to
monitor the spatial distribution of hexachlorocyclohexanes (HCH) and
dichlorodiphenyltrichloroethane (DDT). However, the concentration capacity
and sampling rate of plant tissues vary with the plant species and age, the
sampling location and season (as shown, for example, by Ockenden et al.,
1998b; Müller et al., 2001; Kylin and Sjödin, 2003; Hellström et al., 2004;
Lehndorff and Schwark, 2004). For instance, Müller et al. (2001) found
interspecies differences between leaves and grasses that were attributed due to
differences in the size of the lipophilic compartments and the ratio between
the surface area and the volume of the plant species. Ockenden et al. (1998b)
also found that nonpolar aromatic compounds were accumulated to a higher
degree in lichen than in pine needles. Thus, these kinds of plant data were
subject to several uncertainties, which complicate their interpretation and
limit the potential of plants as monitoring tools.
Thereafter, monitoring of nonpolar aromatic compounds has been achieved
by man-made PAS. Even though man-made PAS have to be deployed and
the samplers are more costly, they are preferred over plants because their
uniform sampler design facilitates both comparisons between sites and
control of the sampling time. Most work has focused on the use of integrative
(from weeks up to years) PAS with high concentration capacities for POPs
such as SPMDs (e.g., Ockenden et al., 1998a; Ockenden et al., 2001;
Lohmann et al., 2001; Meijer et al., 2003; Bartkow et al., 2004), polyurethane
foam (PUF) disks (e.g., Shoeib and Harner, 2002; Jaward et al., 2004a; 2004b),
XAD-2 resin-based passive samplers (e.g., Wania et al., 2003), polymer-coated
glass samplers (POGs) (e.g., Harner et al., 2003) and different leaveresembling samplers with lipophilic coated surfaces like, for example tristerincoated fiberglass sheets (Müller et al., 2000). Monitoring of the continental
atmospheric distribution of nonpolar aromatic compounds using PAS has
8
2. Passive air sampling
only recently been achieved. In 2004, for instance, PUF disks were used to
monitor the spatial atmospheric distribution of PCBs, polybrominated
diphenyl ethers (PBDEs), organochlorine pesticides (OC), polychlorinated
naphthalenes and PAHs simultaneously in 22 countries across Europe
(Jaward et al., 2004a; 2004b). Furthermore, the spatial variation in PCBs and
hexachlorobenzene (HCB) in Norway and the United Kingdom (UK) was
measured by the deployment of SPMDs at 10 and 11 sites for two years
between 1994 and 1996 (Ockenden et al., 1998c), and 1998 and 2000 (Meijer
et al., 2003), respectively.
2.2 Passive sampling using semipermeable membrane devices
(SPMDs)
Semipermeable membrane devices (SPMDs) were developed by scientists at
the United States Geological Survey’s (USGS), Columbia environmental
research center (CERC) as integrative, passive, in situ samplers of the
dissolved fraction of moderately to highly hydrophobic organic compounds
in water (US Patents, 5,098,573, Huckins et al., 1992; 5,395,426, Huckins et
al., 1995). Thereafter, Petty et al. (1993) demonstrated in a laboratory study
that SPMDs are also highly efficient passive samplers of gaseous nonpolar
organic pollutants in air. In a study by Prest et al. (1995) the application of
SPMDs as PAS in field was demonstrated.
The SPMD was developed to mimic the bioconcentration process of
hydrophobic organic pollutants including diffusion through a cell membrane
and accumulation into the lipid tissues of an organism. Thus, SPMDs can be
viewed as an intermediate between conventional sampling techniques and
biomonitoring methods (collecting aquatic organisms) which have
traditionally been used to assess environmental exposure (Figure 3). In
exposure assessments, the identification of new and existing pollutants, and
the determination of TWA pollutant concentrations (and thus estimation of
the environmental exposure over time), are fundamental steps. The SPMD
technique has the advantages over traditionally used grab water sampling
methods like liquid-liquid extraction (LLE), solid phase microextraction
(SPME) and solid phase extraction (SPE), that these steps can be facilitated
without multiple samples having to be collected over time. SPMDs also
facilitate the detection of trace and ultra-trace levels of POPs, because the
integrative sampling approach can pre-concentrate the pollutants in situ to
levels above detection limits. Utvik et al. (1999), for instance, compared the
9
2. Passive air sampling
five sampling methods based on SPMDs, mussels, an in situ large volume SPE
sampler, a SPE disk and LLE, respectively, for monitoring PAH
concentrations in seawater (in this study the North Sea) and SPMDs and
mussels were found to be the most appropriate approaches for analysing
PAHs in seawater.
Figure 3. Picture of a standard 1-mL triolein SPMD placed in water.
A standard SPMD, used for sampling, consists of a 91.4 cm long, 2.5 cm wide
and 75-90 µm thick (wall thickness) layflat non-porous tube (tube with no
fixed pores, only transient cavities with a diameter ∼ 10 Å) of low-density
polyethylene (LDPE) filled with 1 mL (0.915 g) of > 95 % pure triolein (1,2,3tri[cis-9-octadecenoyl] glycerol) (Figure 4). A standard 1-mL SPMD has a total
mass of 4.55 g with a 4:1 LDPE membrane to triolein (w/w) ratio and a
membrane surface-area to triolein volume ratio of ∼ 450 cm2 ⋅ mL-1. The ends
of the LDPE tube have loops to facilitate its deployment during sampling
(Figure 4). In early work, the use of low-molecular-weight dialysis cellulose
membrane filled with organic solvent (hexane) was evaluated (e.g., Södergren,
1987; Herve et al., 1991; Johnson, 1991). Although these devices had several
drawbacks, field studies showed that the in situ passive sampling approach had
potential (Södergren, 1987; Herve et al., 1991; Johnson, 1991). In further
research, the use of the nonporous polymers LDPE, polypropylene, polyvinyl
chloride, polyacetate and silicone or silastic as membranes were evaluated
according to a number of criteria including, for instance, analyte uptake rate,
dialytic performance and cost (Huckins et al., 2002b). Of the polymers tested
only LDPE and polypropylene were found to be acceptable (Huckins et al.,
2002b). LDPE tubes were chosen as membranes because their transient
cavities have similar sizes (∼ 10 Å) to the size limits for particles that can be
transported through biomembranes (9.8 Å) (Opperhuizen et al., 1985). Thus,
10
2. Passive air sampling
only the truly dissolved organic pollutants with molecular masses < 600
approximately Daltons can diffuse into the membrane tube of the SPMDs, as
confirmed for SPMDs in water by Ellis et al. (1995). Triolein is used as the
lipid compartment of the SPMD because it is the major storage fat found in
aquatic organisms, and the triolein-water partition coefficient (KTW) and the
octanol-water partition coefficient (KOW) (a property describing the solubility
in water compared to octanol (a lipid phase)) of hydrophobic, organic
compounds are very similar (Chiou, 1985). However, SPMDs have a much
higher lipid content than aquatic organisms (20 percent in SPMD vs 1-10
percent in fish) and the membrane tube has a higher capacity to concentrate
hydrophobic compounds than biomembranes. Thus, when comparing
SPMDs and tissues of organisms on a mass-basis, (for example, 1 g SPMD
versus 1 g fish), SPMD has a higher concentration capacity for hydrophobic
pollutants. SPMDs can therefore be deployed for relatively short time
periods, from days to weeks, and still ensure that detectable levels are
obtained in the SPMDs (Huckins et al., 2002b). Ockenden et al. (2001), for
instance recommended minimum sampling times for PCBs in air of 3-20
days, depending on the congeners present, the limit of detection and the air
concentrations.
Transient
caves
size < 10 Å
AIR
AIR
Triolein
(lipid)
Organic
pollutant
Membrane
75-90 mm
thickness
Low density
polyethylene (LDPE)
membrane tube
91.4 cm long
2.5 cm wide
LDPE
loop
Figure 4. Schematic diagram of a standard 1-mL triolein SPMD including a thin film
of triolein (neutral lipid found in many aquatic organism) heat sealed inside a lay flat
thin-walled tube of non-porous low-density polyethylene (LDPE). The ends of the
LDPE tube have loops to facilitate the deployment during sampling.
11
2. Passive air sampling
2.2.1 Sampling principles and factors influencing the sampling rate
M at time t depends on air concentration, sampled
compound, sampler design (SPMD configuration +
protective devices) and environmental conditions.
M at equilibrium depends on air concentration, sampled
compound and SPMD configuration (gives KSPMD).
Equlibrium
C
Li n
ea
r
Amount M in SPMD (ng)
In a SPMD, passive sampling is achieved by physical absorption in the
membrane tube, or diffusion through the membrane tube resulting in
absorption of the analyte(s) in the triolein. The uptake of a compound in a
SPMD occurs in three different phases; the linear, curvilinear and equilibrium
phases (Figure 5). The amount taken up (M) by a SPMD depends on (i) the
physicochemical properties of the compound sampled, (ii) the ambient air
concentration, (iii) the exposure time, (iv) the sampler design and (v) the
environmental sampling conditions. The sampling rate (RS), and the time a
chemical remains in the linear and the curvilinear phases until it reaches
equilibrium, are independent of the ambient concentrations but depend on (i)
the design of the sampler, (ii) the pollutants at the sampling site and (iii) the
environmental conditions. Shoeib and Harner (2002) for instance, reported
that PCB 28 (tri-CBs) and PCB 52 (tetra-CB) reached equilibrium after 50-75
and 200-400 days, respectively, while the uptake of PCBs with octanol-air
partition coefficient (KOA) values > 9 like PCBs 101 (penta-CB), 137 and 138
(hexa-CBs) was linear during the entire 450-day sampling.
ear
ilin
urv
t in each phase is independent of the air concentration
but depends on sampled compound, sampler design
(SPMD configuration + protective devices),
environmental conditions. (gives RS, ku, ke)
Exposure time t (day)
Figure 5. The three mass uptake phases that occurs in a SPMD as a function of the
exposure time t in days. The time in each phase depends on the physicochemical
properties of the compound sampled, the sampler design including the configuration
of the SPMD and the devices protecting the SPMDs, and the environmental
conditions (modified from Huckins et al. (2002b)).
12
2. Passive air sampling
Typically, SPMD sampling is performed either in the linear and timeintegrative uptake phase or at equilibrium (Figure 5). In the latter sampling
approach, equilibrium ambient concentrations are estimated and the method
has the advantage that site effects of the environmental conditions are
negligible in some cases. However, the time to approach/reach equilibrium
and the magnitude of the SPMD equilibrium partition coefficient (KSPMD) also
depend on the environmental conditions, and the user has to demonstrate
that steady state has been achieved. This sampling approach is mainly
recommended when the sampling situation is quite constant, like in indoor
environments. However, the in situ integrative sampling approach is
recommended in most cases and was used in the work underlying this thesis.
As mentioned previously, there are several advantages with linear and
integrative sampling. Episodic pollutant releases are more easily detected,
TWA concentrations (which are required in exposure assessments for
organisms) can be estimated, and pollutants present in trace or ultra-trace
levels are concentrated to levels above detection limits. However, this
sampling approach is affected by the environmental conditions at the
sampling site.
The uptake in the SPMD can be restricted by three possible barriers; a thin
boundary or diffusion layer at the membrane-exposure medium interface, the
membrane and the lipid phase (triolein) of the SPMD. The diffusive mass
transfer in the boundary layer is controlled by diffusion, whereas the mass
transfer in the membrane, due to the transient cavity-structure of the
membrane, is restricted by both diffusion and permeation. The transient
cavities of the membrane restrict the passage of all compounds, but the truly
dissolved or gaseous compounds, as mentioned earlier, will be able to
permeate through the membrane, and accumulate in the SPMDs. Thus, only
the TWA concentrations of the primary bioavailable pollutants will be
estimated from the amounts found in the SPMDs. The resistance to diffusive
mass transfer in the lipid of the SPMD is small in comparison to the
resistance in the membrane and the aqueous boundary layer, and can be
neglected. In water, biofouling at the exterior membrane surface can also
restrict the uptake in the SPMDs. However, the uptake is mainly controlled
by the boundary layer or the membrane.
In water, several studies have indicated that the linear and integrative uptake
of hydrophobic compounds (log KOW > 4.4), is mainly controlled by the
boundary layer (Booij et al., 1998; Huckins et al., 1999; Vrana and
Schuurmann, 2002). Thus, the thickness of the boundary layer, and therefore
13
2. Passive air sampling
the water flow/turbulence, effect the uptake of most POPs in SPMDs from
water (Booij et al., 1998; Vrana and Schuurmann, 2002). For compounds with
log KOW < 4.4, the SPMD-water partition coefficient (KSPMD) decreases so that
predominantly the membrane controls their uptake. The uptake in the
membrane depends on the compounds physical/chemical properties, like
molecular three-dimensional geometry and functional groups, because of
possible polymer-molecular interactions and steric effects in the membrane.
Thus, the advantage of boundary layer control compared with membrane
control is that the diffusive mass transfer is non-specific for compounds with
similar size, and hydrophobicity, and in air, also volatility (KOW vs KOA). For
instance, the water diffusion coefficients of compounds with molecular
masses up to 500 g ⋅ mol-1 are in the order of 10-5 cm2 ⋅ s-1 (Booij et al., 1998).
However, the effect of environmental conditions is important to calibrate and
control. In water, the environmental variables temperature (e.g., Huckins et
al., 1999; 2002a; 2002b; Booij et al., 2003) and biofouling (e.g., Huckins et al.,
1990; 2002a; 2002b) also influence the uptake of compounds in SPMDs.
Temperature (Ockenden et al., 1998a) and high wind-speed/turbulence
(Papers II, III) have been shown to affect the uptake in SPMD from air as
well. However, the effect of biofouling is expected to be negligible in air,
whereas chemicals bound to particles or aerosols can be trapped on the
membrane surfaces and influence the amounts taken up by SPMDs from air
(Lohmann et al., 2001; Bartkow et al., 2004). UV sunlight is another
environmental variable that should be considered during air sampling because
photodegradation of UV-sensitive compounds can occur in the SPMDs
(Orazio et al., 2002; Bartkow et al., 2004).
In air, the temporal and geographical variations over time in temperature,
UV-radiation and wind-speed/turbulence can be high. In particular, the
ambient temperature can vary significantly during the course of a day and
between days. One way to reduce the site effects of environmental conditions
during sampling is to protect the SPMDs from direct exposure to sunlight,
rain, wind and particle deposition by using shelters. Results in Paper III, and
research by Ockenden et al. (2001) have shown that the effect of windspeed/turbulence can be reduced by shelter the SPMDs with metal devices
that allow air to pass freely around the SPMDs. Furthermore, use of
performance reference compounds (PRCs) to assess the biofouling effect on
the uptake in SPMDs from water was proposed by Huckins et al. (1993).
PRCs can be described as analytically non-interfering organic compounds that
exhibit moderately to relatively high SPMD affinities and are added to the
lipid-phase of the SPMD prior to membrane enclosure. (It is critical for the
14
2. Passive air sampling
amount of PRC to be equally distributed in the triolein in order to use the
whole surface area in the release process.) The theory of the PRC approach
holds that the release rates of PRCs are related to the uptake rates of the
analogous native compounds sampled and thus can be used to assess the site
effects of environmental variables (Figure 6). Several studies has shown that
PRCs can also be used to assess the effect of temperature and water
flow/turbulence on the uptake in SPMDs from water (e.g., Booij et al., 1998;
Huckins et al., 2002a; Vrana and Schuurmann, 2002; Booij et al., 2003). Thus,
the use of PRCs will improve the precision of the SPMD sampling. The PRC
approach has also been tested for SPMD sampling in air. In the work
reported in Papers II and III deuterated (2H) PAHs and carbon 13-labelled
(13C) PCBs were used, respectively, as PRCs and it was shown that the release
of these compounds can be used to assess the effect of wind on the uptake in
SPMDs from air. Ockenden et al. (2001) have also used 13C-PCBs and found
a correlation between the uptake and release rates in SPMDs from air. In
addition, Bartkow et al. (2004) used 2H-anthracene and 2H-pyrene as PRCs.
High wind
PRC amount M
Native amount M
Low wind
Exposure time t (day)
Exposure time t (day)
Figure 6. Graphs illustrating how the released amount of performance reference
compounds (PRCs) can be used to assess the effects of environmental variables on
the uptake of the analogous native compounds. (Here assessment of the effect of
variable wind-speed that cause changes in the boundary layer thickness, and thus
affect the uptake of target analytes)
2.2.2 Target compounds
The SPMDs concentrate trace and ultra trace levels of the dissolved and
vapour phases of nearly all nonpolar organic compounds with molecular
weights < approximately 600 Daltons and a cross sectional diameter <
15
2. Passive air sampling
approximately 10 Å, to levels that are much higher than the ambient
concentrations and hence detectable. Thus, most nonpolar aromatic
compounds can be sampled by SPMDs from the environment. Listed below
are examples of compound classes that have been taken up in SPMDs from
water (Huckins et al., 2002b; *Ikonomou et al., 2002):
•
•
•
•
•
•
•
•
•
•
•
•
Alkylated phenols (e.g., nonyl phenol)
Chlorinated anisoles and veratroles
Certain heterocyclic aromatics
Chlorinated and brominated benzenes
Organophosphate pesticides
Non-ionic organometallic chemicals
Organochlorine pesticides (OCs)
Polycyclic aromatic hydrocarbons (PAHs)
Polychlorinated biphenyls (PCBs)
Polychlorinated dioxins (PCDDs) and furans (PCDFs)
Polybrominated diphenyl ethers (PBDEs)*
Pyrethroid insecticides
In water, the RS of compounds increases with increasing log KOW and is
maximal around five and six (e.g., Huckins et al., 1999; Luellen and Shea,
2002; Booij et al., 2003). However, for compounds with log KOW < 3, which
are more hydrophilic, the concentration capacity of SPMDs decreases
dramatically as log KOW decreases, and SPMD technique have no advantages
over grab sampling and other relatively low volume techniques. For instance,
Vrana et al. (2002) suggested that the linear and integrative uptake of
compounds under membrane control (log KOW < 4.4) can be as low as two
days in water.
2.2.3 Fields of application
The SPMD technique can be used in a number of different applications to
assess both the ambient levels (and thus environmental exposure) and the
toxicity of organic pollutants. Specific applications in which SPMDs can be
used in field studies are described below.
I. Determination of (i) the presence of new and existing organic
pollutants, (ii) the sources, and (iii) the local, regional and continental
16
2. Passive air sampling
spatial distribution (often in relative terms that is semi-quantitatively)
of pollutants.
II. Estimation of the TWA concentrations of pollutants that are
dissolved or in the vapour phase (quantitatively).
III. Estimation of organisms’ exposure to organic pollutants that are
dissolved or in the vapour phase.
IV. Identification of chemical fractions, with in situ sampled
concentrations, of primary bioavailable organic pollutants that can be
used in bioassays/biomarker tests and immunoassays.
Table 1 gives examples of field studies which have used SPMDs for different
purposes. Many of the applications are closely related, and thus, most of the
listed studies include more than one of the applications. The sampling
environments (water, air and/or soil) of each study are specified. In the Table,
several field studies are listed as having quantitatively determined the TWA
concentrations. However, in some of these studies, the sampling rates used
were estimated and their approach was semi-quantitative rather than truly
quantitative.
17
2. Passive air sampling
Table 1. Examples of field studies which have used SPMDs for different purposes and in different environments.
*Purposes of the field studies presented in the work underlying this thesis.
Purpose
Field studies
(sampling environment in brackets)
Ii: To determine presence*
Lebo et al., 1992 (creek); Prest et al., 1992 (river); Prest et al., 1995
(seawater+air); Strandberg et al., 1997 (compost); Bergqvist et al., 1998
(seawater); Ockenden et al., 1998a (air); Ockenden et al., 1998c (air); Echols
et al., 2000 (river); Zimmerman et al., 2000 (rivers); Rantalainen et al., 2000
(soil at lake shore); Sabaliunas et al., 2000 (rivers); Granmo et al., 2000
(seawater); Booij and van Drooge, 2001 (seawater+air); Lohmann et al., 2001
(air); Lindström et al., 2002 (lakes), Paper I, 2003 (air); Isidori et al., 2003 (air);
Meijer et al., 2003 (air); Rastall et al., 2004 (lake); Balmer et al., 2004 (lakes);
Petty et al., 2004 (river+waste water); Paper V, 2004 (indoor air); Shaw et al.,
2004 (river+seawater); Paper IV, 2004 (air).
Iii: To determine sources*
Iiii: To determine the spatial
distributions
Local*
Regional*
Continental*
Granmo et al., 2000 (seawater); Lohmann et al., 2001 (air); Lindström et al.,
2002 (lakes); Meijer et al., 2003 (air); Paper I, 2003 (air); Balmer et al., 2004
(lakes); Petty et al., 2004 (river+waste water); Paper V, 2004 (indoor air);
Shaw et al., 2004 (river+seawater); Paper IV, 2004 (air).
Prest et al., 1992 (river); Prest et al., 1995 (seawater+air); Bergqvist et al., 1998
(seawater); Echols et al., 2000 (river); Rantalainen et al., 2000 (soil+lake
shore); Granmo et al., 2000 (seawater); Booij and van Drooge, 2001
(seawater+air); Rastall et al., 2004 (lake); Paper V, 2004 (indoor air).
Sabaliunas et al., 2000 (rivers); Zimmerman et al., 2000 (rivers); Lohmann et
al., 2001 (air); Lindström et al., 2002 (lakes), Paper I, 2003 (air); Balmer et al.,
2004 (lakes); Shaw et al., 2004 (river+seawater).
Ockenden et al., 1998c (air); Meijer et al., 2003 (air); Paper IV, 2004 (air).
II: To estimate time weighted
average (TWA) concentrations
of pollutants*
Prest et al., 1995 (seawater+air); Bergqvist et al., 1998 (seawater); Ockenden
et al., 1998a (air); Ockenden et al., 1998c (air); Echols et al., 2000 (river),
Rantalainen et al., 2000 (soil at lake shore); Granmo et al., 2000 (seawater);
Lindström et al., 2002 (lakes); Paper I, 2003 (air); Balmer et al., 2004 (lakes);
Petty et al., 2004 (river+waste water); Paper V, 2004 (indoor air); Shaw et al.,
2004 (river+seawater).
III: To estimate
exposure*
organisms’
Echols et al., 2000 (river); Zimmerman et al., 2000 (rivers); Paper I, 2003 (air).
IV: To identify chemical
fractions, with in situ sampled
concentrations,
of
primary
bioavailable organic pollutants
that
can
be
used
in
bioassays/biomarker tests and
immunoassays
Sabaliunas et al., 2000 (rivers); Isidori et al., 2003 (air); Rastall et al., 2004
(lake).
2.2.4 Theory and mathematical models of SPMD sampling
The basic theory of SPMD sampling, and mathematical models used to
estimate the water ambient concentrations from the SPMD concentrations,
were developed by Huckins et al. (1993). In the early work, triolein (lipid) was
considered to be the only compound-accumulating compartment of the
18
2. Passive air sampling
SPMD. The concentration of a compound in the lipid compartment of the
SPMD is given by (Huckins et al., 1999)
CL = Cw · KLw(1-exp[-ko · Kmw· A· t · KLw-1· VL-1])
Eq. 4
where CL is the concentration in the SPMD lipid compartment, Cw is the
water concentration, KLw is the lipid-water partition coefficient, ko is the
overall mass-transfer coefficient, Kmw is the membrane-water partition
coefficient, A is the SPMD surface area, t is the exposure time in days, KLw is
the lipid-water partition coefficient and VL is the lipid volume.
However, the membrane tube also accumulates organic pollutants and, thus,
contributes to the amount found in the SPMD. To include the concentrations
recovered by the membrane and thus, estimate the water concentrations from
the amounts found in the SPMDs, Huckins et al. (1999) suggested that the
membrane can be expressed as a lipid-equivalent volume and, thus, the
SPMD can be considered as a single-compartment model. However, the
differences between the triolein and the membrane tube in both total mass
(4:1 membrane/triolein ratio), and concentration capacity (triolein has about
four times higher capacity (Huckins, 2004)) has to be compensated for in the
model. For compounds under boundary layer control, the uptake in amount ⋅
g SPMD-1, CSPMD, is then described by
CSPMD = Cw · KSPMD(1-exp[-kw · A · t · KLw-1 (VL-1 + KmL-1 · Vm-1)])
Eq. 5
where KSPMD is the SPMD-water partition coefficient, kw is the mass transfer
coefficient in the boundary layer, KmL is the membrane-lipid partition
coefficient, and Vm is the volume of the membrane.
When the uptake of compounds is controlled by the membrane, the pollutant
concentrations in the SPMD are described by
CSPMD = Cw · KSPMD(1-exp[-km · Kmw · A · t · KLw-1 (VL-1 + KmL-1 · Vm-1)]) Eq. 6
where km is the mass transfer coefficient in the membrane.
To simplify eqs. 2 and 3, the terms kwA/KLw(VL+KmLVm) and
kmKmwA/KLw(VL+KmLVm), respectively, in each equation are expressed as the
exchange coefficient, ke (t-1), and the concentration in the SPMD is calculated
according to
19
2. Passive air sampling
CSPMD = Cw · KSPMD(1-exp[-ke · t])
Eq. 7
Under linear and integrative sampling, when the term ket is small or CSPMD/CW
<< KSPMD, the eq. 7 can be reduced to
CSPMD = Cw · ku · t = Cw · RS · t · MSPMD-1
Eq. 8
where ku is the uptake rate constant in mL · d-1· g SPMD-1 for a specific
compound, RS is the volume of water sampled per unit time in dm3 · d-1 at a
given temperature, and MSPMD is the mass of the SPMD in g. The uptake and
the elimination of the SPMD are related according to
ku = KSPMD · ke
Eq. 9
where KSPMD is the SPMD-water partition coefficient.
The basic theory of SPMD sampling in water developed by Huckins et al.
(1993, 1999) was used in this thesis to model the SPMD sampling in air,
assuming that the uptake and elimination processes of SPMDs in water and
air are related. In the work underlying this thesis only standard 1-mL triolein
SPMDs and the integrative sampling approach were used. Thus, the
integrative model as expressed by eq. 8 was used to describe the uptake in
SPMDs. However, some changes to that model were introduced. First, the
subscript SPMD was replaced with SA for SPMD in air. Second, the SPMD
concentration, CSPMD, in g ⋅ g SPMD-1 was not modelled since standardized
SPMDs were used. Instead, the amount, MSA, in ng ⋅ standard SPMD-1 was
modelled and thus, the term MSPMD was not included in the equation, giving
MSA= CA · ku · t = CA · RS · t
Eq. 10
where CA is the air concentration in ng · m-3, ku is the uptake rate constant in
mL · d-1 for a standard 1-mL SPMD, and RS is the volume of air sampled per
unit time in m3 · d-1, at a given temperature, for a standard 1mL SPMD. In
cases where standardized SPMDs are not used, the ku and the RS values have
to be normalized by the differences between the masses of the SPMD
configuration used and a standard 1-mL SPMD.
20
2. Passive air sampling
2.2.4.1 The PRC model
As mention earlier, PRCs can be used to calibrate the RS of SPMD in situ and
thus assess the effect of environmental variables (Huckins et al., 1993). Two
types of PRC approaches can be used to assess the differences between the
laboratory pre-calibrated RS and the RS values in situ. One approach is to use
PRC-derived exposure adjustment factors (EAFs), described in detail by, for
instance, Huckins et al. (2002a). The second approach, described in detail by
for instance, Booij et al. (1998), was used in the work underlying this thesis,
and is based on the use of ke values calculated from the release of PRC
according to
MPRC-A = M0PRC-A exp(-ke · t)
Eq. 11
where MPRC-A is the amount of the PRC in the SPMD at time t and M0PRC-A is
the amount of the PRC in the SPMD at the beginning of the sampling. If the
amount of PRC in the SPMD is measured at the beginning and end of the
sampling, as was done in work underlying this thesis, eq. 11 is solved as a
two-point derivation of ke (Huckins et al., 2002b)
ke = ln(M0PRC-A/MPRC-A) · t-1
Eq. 12
When the uptake in the SPMD is linear and integrative, the estimated ke value
can be used in eq. 8 and the air concentration of the sampled compound is
given by
CA = MSA · ke-1 · KSA-1 · t-1
Eq. 13
where KSA is the SPMD-air partition coefficient.
When the PRC amount released is ≥ approximately 60 percent, the uptake of
compounds with similar or lower SPMD affinity is nonlinear or at
equilibrium. Gale (1998), for instance, suggested that at 60 percent of the
steady state concentrations the uptake will be in the curvilinear phase. In such
cases, use of eq. 13 will underestimate the CA and instead an exponential or
equilibrium model should be used.
In summary, for quantitative data interpretation (converting amounts found
in the SPMDs to air concentrations) relevant RS (see eq. 10) or KSA (see eq.
21
2. Passive air sampling
13) values have to be known. However, for the time being no KSA values are
available, and the KOA values have to be used as surrogates for the KSA values
as were suggested by, for instance, Ockenden et al. (1998a).
22
3. Compounds sampled and their physical/chemical properties
3. COMPOUNDS SAMPLED AND THEIR PHYSICAL/CHEMICAL
PROPERTIES
The work underlying this thesis was focused on the nonpolar aromatic
compounds PCBs, PAHs, alkyl-PAHs and nitro-PAHs. Below selected
physical/chemical properties of these compounds are presented.
3.1 Polycyclic aromatic hydrocarbons (PAHs)
PAHs are a wide class of compounds which consist of different numbers of
fused benzene rings (generally two to six) in linear, angular or cluster
arrangements (Figure 7). The main concerns of PAHs are their wide
distribution in the environment, their persistence and the fact that some
PAHs have reported mutagenic and carcinogenic properties associated with
increasing size of the molecule (compounds with four or more benzene rings
being especially carcinogenic), and their metabolic transformation to reactive
dihydrodiol epoxides (Pickering, 1999). They are present in the atmosphere
primarily due to emissions from gasoline- and diesel-powered motor vehicles,
municipal and commercial incinerators, residential heating systems that
combust fuels like coal, wood, gas and oil, various industrial processes and
volatilization from polluted soils. Over 500 PAHs have been detected in air,
but most studies focus on the 16 PAHs designated as priority pollutants by
the United States Environmental Protection Agency (the 16 EPA PAHs).
PAHs can also exist in substituted forms like alkyl- or nitro-PAHs (Figure 7).
High concentrations of alkyl-PAHs compared to unsubstituted PAHs have
been found in emissions of petrogenic sources like traffic and fossil fuel
combustion systems for home heating (see, for instance, Ramdahl, 1983;
Rogge et al., 1993; Nielsen, 1996; Lim et al., 1999; McDonald et al., 2000).
Nitro-PAHs originate from both primary sources like emissions of diesel
motor vehicles (e.g., Nielsen, 1984; Feilberg et al., 1999; Bamford et al., 2003),
and reactions of the parent PAHs in the atmosphere (e.g. Nielsen, 1984; Arey
et al., 1986; Atkinson et al., 1987; Atkinson and Arey, 1994). The main
concerns regarding nitro-PAHs are that many of them are reported to have
direct mutagenic activity and carcinogenicity (e.g., Tokiwa et al., 1994; Durant
et al., 1996).
23
3. Compounds sampled and their physical/chemical properties
a)
b)
c)
Figure 7. Molecular structures of a) the EPA PAH phenanthrene, b) 1methylphenanthrene and c) 2-nitrofluorene.
To our knowledge, the KOA of PAHs have only been measured in one study
by Harner and Bidleman (1998), who determined the log KOA values of
fluorene, phenanthrene, pyrene and fluoranthene, respectively at different
temperatures (examples in Table 2, pp 55). Lohmann et al. (2000) suggested
that the log KOA values for 2-ring to 6-ring PAHs ranged between 6 and 12.
Furthermore, PAHs display a wide range of gas-particle partitioning
characteristics in the atmosphere. PAHs with two and three rings, which
have relatively low log KOA values, are mainly associated with the gas phase, 4ring PAHs occur in both the gas and particle phase, while PAHs with higher
log KOA values, such as PAHs with five rings or more, are mainly bound to
particles (see, for instance, Lohmann et al., 2000). In comparison, both alkyland nitro-PAHs appear to have lower vapour pressure than unsubstituted
PAHs with corresponding numbers of benzene rings in their molecular
structure (manifested, for example, by longer retention times in GC-analysis).
For instance, Dimashki et al. (2000) found that 2-ring nitro-PAHs
(nitronaphthalenes) occurred only in the gas phase, nitro-PAHs with three
rings (9-nitroanthracene) were associated with both the gas and particle
phases, whereas nitro-PAHs with four rings or more (1-nitropyrene, 2nitrofluoranthene and 7-nitrobenz[a]anthracene) were mainly bound to
particles.
3.2 Polychlorinated biphenyls (PCBs)
PCBs are a group of compounds consisting of 209 congeners with varying
degrees of chlorination of the biphenyl scaffold: 1 to 10 chlorine atoms in the
molecular structure (Figure 8). In all, there are ten homolog groups (mono- to
deca-CBs). In the work underlying this thesis, the PCBs were numbered
according to the IUPAC rules, a system developed by Ballschmitter and
Zeller (1980). PCBs have been widely used, for instance, as plasticizers, fire
retardants, sealing liquids and dielectric fluids for capacitors and transformers.
They can be found in all environmental compartments including humans, and
24
3. Compounds sampled and their physical/chemical properties
have shown multiple toxic effects on wildlife and humans (e.g., Giesy and
Kannan, 1998; van den Berg et al., 1998; Bosveld et al., 2000; Kannan et al.,
2000; Letcher et al., 2002). The major source of PCBs in the atmosphere is
volatilization from contaminated sites where they have been stored or
dumped, and combustion of PCB-containing material (e.g., Simcik et al.,
1997).
Cln
Figure 8. Schematic diagram of the general molecular structure, substitution
positions, and nomenclature of the PCBs.
Harner and Bidleman (1996) have measured the log KOA values of 15 PCBs
(mono, tetra-, penta-, hexa- and hepta-CBs), at different temperatures, and
found log KOA values between 7 and 11 at 20 °C, that increased with
increasing chlorination (examples in Table 2, pp 55). Within the same
homolog group, PCBs with fewer chlorine atoms in ortho-position were found
to have higher KOA values. Furthermore, Chen et al. (2002) predicted the log
KOA values of the 209 PCBs, at the same temperature, to be between 5.6 and
12.3 (examples in Table 2, pp 55). In the atmosphere, PCBs predominantly
exist in the gas phase (e.g., Ockenden et al., 1998a; Yeo et al., 2003; Tasdemir
et al., 2004). For instance, Yeo et al. (2003) found that 90 percent of the total
PCBs (gas+particle phase) reside in the gas phase, with mono- to hexa-CBs
mainly associated with the gas phase, whereas hepta- to deca-CBs occur in
both the gas and particle phase.
25
26
4. Sampling and chemical analysis
4. SAMPLING AND CHEMICAL ANALYSIS
The design of the sampler and the sampling procedures used in the work
underlying this thesis is presented in detail below. In addition, the chemical
analysis of the SPMD samples, the quality assurance and quality control of the
data, and the data interpretation are described.
4.1 Design of the sampler
The design of the Exposmeter air sampler used in the work described in this
thesis included two standard 1-mL SPMDs obtained from Exposmeter AB,
Tavelsjö, Sweden. Each SPMD was mounted carefully between five steel rods
attached to a 150 × 140 mm steel disc (a steel device called a spider) and the
loops at the ends of the membrane were hung over the two outermost steel
rods (Figure 9). The two SPMDs, mounted on separate spiders, were placed
horizontally on top of each other inside the upper part of a metal umbrella
that was hung with the opening, covered by a metal mesh (340 mm i.d.),
facing downwards. The metal umbrella had an inner diameter of 320 and 160
mm at the bottom and top, respectively, decreasing progressively from its
lower to its upper part. This design of the metal umbrella was chosen to
protect the SPMDs from direct wind, sunlight, rain, and particle deposition
while allowing air to pass under and around the SPMDs.
Figure 9. Design of the Exposmeter air sampler used in the work underlying this
thesis. Two standard 1-mL SPMDs were deployed on separate 150 × 140 mm steel
spiders and placed horizontally on top of each other inside the upper part of a metal
umbrella (320/160 mm i.d.) that was hung with the opening, covered by a metal
mesh (340 mm i.d), facing downwards (Paper III).
27
4. Sampling and chemical analysis
4.2 Sampling procedure and approach
In the work described in this thesis, unexposed SPMDs were delivered from
Exposmeter in gas-tight solvent-cleaned tin cans (five SPMDs per can) and
then stored at -18 °C. The SPMDs were transported in these tin cans to the
sampling locations where the SPMD samplers were set up according to
previous description (4.1). The samplers were hang on metal racks at 1 to 3 m
heights for approximately 21 days when the SPMDs were retrieved and
placed in separate pre-cleaned gas-tight tin cans at -18 °C. The average
sampling temperature at each sampling location was measured in parallel. In
the study reported in Paper IV, environmental conditions such as wind-speed
and wind direction were also measured, when possible. To provide controls
for the purity of the SPMDs used and the contribution of the sampling and
analytical procedures to the amounts found in the SPMDs, one additional
SPMD (referred to as the field control, FC) was exposed to air only during
deployment and retrieval of the SPMDs, and was analyzed using the same
procedure as the other SPMDs.
Air sampling for < 30 days ensures that uptake of most nonpolar aromatic
compounds in the SPMDs is linear, and thus sampling is time-integrative.
Therefore, the integrative model was used to describe the uptake of pollutants
in the SPMDs (see 2.2.4). However, the uptake of some of the most volatile
PAHs, such as naphthalene, are in the curvilinear phase at the end of a 3week sampling period and the naphthalene amounts found in the SPMDs,
calculated as ng ⋅ SPMD-1 ⋅ d-1, were underestimated. In addition, timeintegrative sampling of these compounds does not occur throughout the
sampling and the ability of SPMD to detect episodic pollutant releases
declines.
4.2.1 The PRC approach
2
H-PAHs with molecular weights no higher than that of deuterated chrysene,
together with 13C di- and tri-CBs are commonly used as PRCs. The
compounds selected as PRCs, should be suitable for the environmental
conditions of the sampling sites and the sampling time in order to ensure that
the PRC losses during the sampling are in an acceptable range. The changes
in the initial PRC amounts should ideally meet the criteria of 1-(CSPMD/C0SPMD)
× 100 % > 3 × coefficient of variance (CV) when the PRC releases < 50 %,
and CSPMD/C0SPMD × 100 % > 3 × CV when the PRC releases > 50 % (Huckins
28
4. Sampling and chemical analysis
et al., 2002b). These criteria are generally met when the amount of PRC
released is in the range 20 to 80 % of the initial amount in the SPMDs
(Huckins et al., 2002b). For example, if the sampling time is about two weeks,
2
H-naphthalene will be completely lost, while no loss of 13C-PCB 180 will be
detected. To control the amount of PRC in the SPMD at the beginning of the
sampling, one SPMD, referred to as FC, was analyzed with the same
procedure as the other SPMDs.
4.2.2 Cleaning of the sampling equipment
The sampling devices were carefully cleaned before use. First, the steel spiders
were put in 40 %-hydrochloric acid for 5 minutes and then in a cold water
bath for 10 minutes. The acid-washed steel spiders were brushed and washed
by hand in water for 1-2 minutes, and then heated for 10 hours at 500 °C.
The metal umbrellas were washed by hand in clean water and detergents.
Before retrieval of air exposed SPMDs, the interior of new tin cans were
shaken in 10 mL of acetone for 20-30 s, and then in 10 mL hexane for
another 20-30 s
4.3 Extraction, cleanup and analysis
A wide range of nonpolar organic compounds are taken up in the SPMDs.
Therefore an analytical method that is non-destructive and allows the analysis
of a large number of organic pollutants is often used (see, for instance, Lebo
et al., 1992). The general procedure used for cleaning up SPMD samples,
including the extraction and analysis of target or unknown analytes, is
presented in Figure 10. A detailed description of the quality of the chemicals
used, their sources and the analytical method used in the work underlying this
thesis is given in Paper I. Briefly, the SPMDs were brushed and washed by
hand in water, then shaken in hexane followed by hydrochloric acid and
finally washed in water again, to clean the membrane surface from lipids, soot
and other particles. The externally cleaned SPMDs were checked for holes
and then the organic pollutants were extracted from them by dialysis for 24×2
h in 95:5 (v:v) cyclopentane:dichloromethane, with an exchange of solvent
after 24 h. The extracted SPMDs were discharged and the dialysates were
cleaned with a gel permeation chromatography (GPC) system that included a
high resolution (HR)-GPC column using 65:45 (v:v) hexane:dichloromethane
as eluents. To remove residues of plastic films (polyethylene) and lipids, that
29
4. Sampling and chemical analysis
still interfered with the analysis after the GPC-run, a second clean-up step was
developed by testing various open liquid chromatography systems with a
range of materials, including acid/basic silica, florisil, deactivated silica and a
mixed silica including deactivated silica and basic silicate, with hexane and/or
hexane:dichloromethane (50:50 or 75:25 (v:v)) as eluents (for all of the
systems tested). The mixed silica gel column with 50:50 (v:v)
hexane:dichloromethane as eluent was selected for use in further studies since
it gave acceptable recoveries and high purities of the target analytes (PAHs).
The SPMD samples were analyzed for PAHs, alkyl-PAHs and PCBs by highresolution gas chromatography/low-resolution mass spectrometry
(HRGC/LRMS) operated in electron ionization mode monitoring selected
ions. In Paper IV, the SPMD samples were analyzed for the nitro-PAHs by
HRGC/LRMS operated in negative ion chemical ionization mode, using an
analytical method described in detail by Dusek et al. (2002).
Size exclusion
chromatography
Chemical fraction (mixed silica)*
Organic solvent dialysis
Chemical analysis (GC/MS)*
Exterior cleaning
Figure 10. General procedures used for cleaning up SPMD samples including the
recovery and analysis of target and unknown analytes taken up in exposed SPMDs.
*The specific methods used in the work underlying this thesis are shown in
parentheses.
4.4 Quality assurance and quality control (QA/QC)
Quality assurance (QA) and quality control (OC) are important to ensure the
reliability of the data obtained. One way to improve the quality of the analysis
is to include the use of labelled standards. In the work underlying this thesis,
different labelled standards were used, depending on the target analytes, and
were all added to the extracts after the dialysis. In the studies reported in
Papers I and IV, six 2H-PAHs (naphthalene, fluorene, anthracene, pyrene,
30
4. Sampling and chemical analysis
chrysene and benzo[k]fluoranthene) were used, in a mixture, as internal
standards for the clean-up. In latter study, two 2H-nitro-PAHs (1-nitropyrene
and 6-nitrochrysene) were also added to the dialysates. In the studies
described in Papers II and III, one mixture of the 13C-PCBs 28, 47, 52, 101,
105, 118, 138, 153, 156, 180, 194 and 209, and another mixture of 2H-PAHs
(acenaphthylene, fluoranthene and benzo[g,h,i]perylene were used in Paper II,
and in Paper III, also naphthalene and anthracene was added to the mixture),
were added as internal standards. In all cases, deuterated dibenzofuran was
used as a recovery standard during the GC/MS-analysis. In cases where the
samples were re-analyzed including an extra GPC-run, one 13C tetra- or pentaCB was used as recovery standard. Acceptable recoveries of these standards
were within 50-120 percent. Obtained data were adjusted using these recovery
values. For detection, the signal to noise ratio should be equal to or greater
than three (S/N ≤ 3). In addition, laboratory blanks (LBs) and FCs consisting
of solvent and SPMDs, respectively, exposed to air during the deployment
and retrieval of the samples, were analyzed in the same manner as the SPMD
samples, and were checked for possible interferences within the same
retention times as the target analytes.
The reproducibility of the SPMDs was evaluated (Paper III) by analysing the
two SPMDs of each sampler separately. The uptake of PAHs and PCBs
between the replicate SPMD samples (derived from 21-day sampling and
chemical analysis) was 17 and 19 percent, respectively. However, differences
in the amounts taken up by some replicate SPMDs were found, indicating
that the extremely high wind-speeds applied in this study could have caused
wind conditions to differ inside the protective metal umbrellas and thus cause
differences in the amounts taken up by the two SPMDs in each sampler.
However, in the work presented in Paper III, we found that most replicate
SPMDs, sequestered and released similar amounts of the gas-phase PAHs and
PCBs, respectively. These results agree with other studies that have concluded
that SPMDs sequester gas phase PCDD/Fs, PAHs and PCBs from the
atmosphere with good reproducibility (e.g., Lohmann et al., 2001).
31
4. Sampling and chemical analysis
4.5 Data interpretation
4.5.1 Semi-quantitative and quantitative approach
In Papers I-IV, the SPMD data were interpreted by comparing differences in
amounts and profiles found in the SPMDs, and were expressed in ng ⋅
SPMD-1 ⋅ day (d)-1. Since the reproducibility (comparability) of the SPMDs
was high, but differences in the RS values between different target analytes
and sampling sites were not accounted for, this approach to interpreting data
has been defined as semi-quantitative. For quantitative data interpretation
(converting amounts taken up in the SPMDs to air concentrations) relevant
RS (see eq. 10) or KSA (see eq. 13) values have to be known.
When Paper I was written, available calibration data for SPMDs in air were
limited to field-calibrated RS values for PCBs reported by Ockenden et al.
(1998a, 2001) and Shoeib and Harner (2002). Therefore, the RS values of
PAHs in water calibrated by Huckins et al. (1999) were used for preliminary
calculations of the PAH concentrations in air. At that time, no adjustments
were made for the differences in density between water (1.0 kg ⋅ L -1) and air
(1.2 kg ⋅ m-3), and the RS values of the PAHs in water expressed in L ⋅ d-1 were
used as RS values in air expressed in m3 ⋅ d-1. Thus, this approach to interpret
data was not quantitative, although the values were corrected for different RS
values for PAHs in water.
In Paper III, the amounts in the SPMDs were converted to air concentrations
for some selected PAHs and PCBs using the in situ-calibrated ke values of 2Hphenanthrene and 13C-PCB 15, respectively, and the eq. 13. For the selected
PAHs and PCBs, no KSA values were available. Ockenden et al. (1998a) have
successfully used the KOA values as surrogates for the KSA and, thus, this
approach was applied in Paper II. The KOA values can be directly measured as
was made by Harner and Bidleman (1996, 1998) (used in Paper III) or
estimated from
KOA = KOW ⋅ KAW-1 = KOW ⋅ R ⋅ T ⋅ H-1
Eq. 14
where KAW is the air-water partition coefficient, H is the Henry’s law constant,
R the gas constant, and T is the temperature (used in Paper II).
32
4. Sampling and chemical analysis
Thus, the approach used in Paper III corrected for the site effects of different
sampling conditions, but since no KSA values were available for each
compound, it was semi-quantitative rather than truly quantitative. To clarify
that the data in Paper III were semi-quantitative, the air concentration data
were normalized by the highest concentration determined for each
compound.
4.5.2 Multivariate data analysis using principal component analysis (PCA)
For interpretation of the data in Paper IV presented in this thesis, principal
component analysis (PCA) was used (Jackson, 1991). In a PCA the variations
in a multivariate data matrix X with n rows (objects) and k columns
(variables) are projected into a few uncorrelated (orthogonal) principal
components (PCs). Basically, the systematic variation of the X matrix is
extracted into smaller matrices, the score T and loading P´ matrices,
extracting the variation between objects and variables respectively. The first
PC is oriented to explain as much variation in the data as possible and
presents the best linear summary of X. The second PC is orthogonal to the
first, and explains the next largest variation in the data, and so forth. In the
multivariate data analysis of this thesis, the number of significant PCs in the
PCA was determined by cross-validation (Wold, 1978).
The PCs define a plane with dimensionality of the number of PCs. Each
object can be projected onto this plane and the coordinates in this plane gives
the score for the object (i.e., the distance between the origin of the plot and
the object’s projection into the plane) (Figure 11a). Thus, the score (t) gives
the importance of the object for the PC. In a score plot, the scores of two
PCs are plotted against each other and thus, visualize the relationship between
objects. Hence, two objects close to each other in a score plot are similar and
objects far apart are not. The direction of each PC is defined by the loadings
(p) as given by the cosine of the angles between the original variables and the
PC (Figure 11b). Hence the p values show how the original variables “load”
into the PC. Thus, by plotting the p values of two PCs against each other, the
importance of each variable for each PC can be revealed. Comparing the
score and the loading plot of two PCs reveal the importance of each original
variable for each object, e.g. which individual PAH concentrations (variables)
that are high or low at the sampling site (object).
33
4. Sampling and chemical analysis
a)
b)
Figure 11. Illustration of the projection of multivariate data into a plane with
dimensionality of a few uncorrelated principal components (PCs). a) The score (t)
(the distance between the origin and the object’s projection into the plane) for each
object gives its importance for each PC, while b) the loadings (p) (the cosine of the
angles between the original variables and the PC) gives the importance of each
variable for each PC (modified from Eriksson et al. (1999)).
PCA was used in Paper IV to reveal whether there were spatial differences in
the concentration of gas phase PAHs within and between five European
countries (Austria, Poland, Slovakia, Sweden and the Czech Republic) and to
discern, if possible, the PAH-patterns of potential sources. PCA has been
performed on SPMD data by Echols et al. (2000) to compare the relative
PCB patterns of SPMDs, caged fish and sediments. Two different PC
analyses were performed on the data set in our study, which consisted of 11
columns (variables) and 79 rows (objects). The object AU2-99 and the two
variables dibenz[a,h]antracene and fluorene were excluded from the two
models since more than 50 percent of the corresponding values in the dataset
were missing. The variables (the individual PAH concentrations in the data
matrix) showed a positive skewness for the objects (sampling sites) when
plotted in a histogram. The data were therefore log-transformed to obtain an
approximately normal distribution. In the first PCA all variables were meancentred and scaled to unit variance, while the data in the second PCA were
mean-centred and normalized by the total PAH amount sampled at each
sampling location. The data were normalized to avoid the results being
affected by the differences in PAH-concentrations between objects. The
software package SIMCA-P 9.0, 2001, obtained from Umetrics AB, SE-907
19 Umeå, Sweden, was used for the multivariate data analysis presented in
work of this thesis.
34
5. Factors influencing the uptake of compounds in SPMDs from air
5. FACTORS INFLUENCING THE UPTAKE OF COMPOUNDS
IN SPMDs FROM AIR
When testing, for instance, the sampling rate, RS, or the effect of air velocity
for diffusive samplers of gases and vapours, European standards require the
tests to be performed under laboratory conditions in an exposure chamber
made of material that is inert with respect to the measured substance
(Anonymous, 2002b). In addition, a constant concentration of a calibration
gas mixture that either can be traceable to national standards or has been
determined by a reference method should be generated in the exposure
chamber. When determining the wind effects, different sampling conditions,
with variations in exposure time, the air velocity and the sampler orientation
should be tested, and at least six samplers should be exposed for each test
condition. The results from the six (or more) test samplers should be
compared with results obtained simultaneously with six samplers of a
reference method or one independently calibrated instrument. When
determining the RS of the target analytes, 24 test samplers should be used
instead of six. The volume of the generated atmosphere (calibration gas),
passing through the exposure chamber, should have sufficient capacity to
accommodate all these samplers simultaneously. In addition, the samplers in
the exposure chamber should be positioned in such a manner ensuring that
there is no interference between each sampler. If the outlet concentration,
determined by a reference method or an independent measurement, differs by
more than 5 percent from the inlet concentration, the generated air volume is
too low or the samplers are positioned too close to one another, and the
exposure system should be modified.
The European standard includes several requirements that are problematic to
fulfill when testing the performance of standard SPMDs as PAS of nonpolar
aromatic compounds. Firstly, generating a calibration gas with constant
concentrations of POPs like PAHs and PCBs that partition both to the gas
phase and to particles involves several difficulties:
I. These compounds have quite low gas phase solubility and generating
a gas of these compounds is complicated.
II. Generating gas concentration that is constant over time is difficult
because the amounts of aerosols and particles in the atmosphere of
the exposure chamber can vary and that will effect the gas
concentrations. Since these compounds display a wide range of gasparticle partitioning characteristics, this effect will also vary between
compounds.
35
5. Factors influencing the uptake of compounds in SPMDs from air
Another requirement that is problematic to fulfill is to generate an air volume
(calibration gas) in the exposure chamber which has the capacity to
accommodate the required SPMD samplers and reference samplers
simultaneously throughout the sampling period while ensuring that the gas
concentrations of the analytes change by less than 5 percent. For instance,
ensuring that the SPMDs have at most a 5 percent effect on the air
concentration during a 21-day continuous air sampling with six SPMD
samplers, and six MeVols, which is a standard method, requires 12,600
(assumed RS: 5 m3 ⋅ d-1) and 35,280 (assumed RS: 14 m3 ⋅ d-1) m3 air,
respectively. Due to these difficulties, the European standard conditions for
laboratory exposure tests are almost impossible to meet when testing the
performance of SPMDs as PAS of nonpolar aromatic compounds.
5.1 Physical/chemical properties of the compounds sampled
The SPMDs concentrate trace and ultra trace levels of the dissolved and
vapour phases of nearly all nonpolar aromatic compounds with molecular
weights < approximately 600 Daltons and a cross sectional diameter <
approximately 10 Å. In water, KSPMD (i.e. the concentration capacity) and the
RS of the compounds, depend on the log KOW value of the target analyte (e.g.,
Huckins et al., 1990; Booij et al., 1998; Meadows et al., 1998; Huckins et al.,
1999; Luellen and Shea, 2002; Vrana and Schuurmann, 2002; Booij et al.,
2003). Luellen and Shea (2002), for instance, suggested RS values of PAHs
and alkyl-PAHs in the range 2.11 to 6.06 L ⋅ d-1 that increased with increasing
log KOW values up to about 5-5.5, where the RS began to decrease, presumably
due to steric effects of the large molecular weight fused-ring PAHs. They also
found that PAHs with log KOW > 4.5 remained in the linear uptake phase
throughout a 30-day sampling period, whereas the uptake of PAHs with log
KOW < 4.0 or between 4.0 and 4.5 was curvilinear after 7 and 13-15 days,
respectively. Furthermore, Booij et al. (2003) suggested that the RS of
chlorobenzenes, PCBs and PAHs, respectively, were all in the range 20-200 L
⋅ d-1 (which are high, probably due to the effect of high water flow), and
increased with increasing log KOW values up to about six when the RS slightly
decreased again. For compounds with log KOW values below three, the
concentration capacity (the KSPMD) of SPMDs decrease dramatically, and the
use of SPMDs as passive water samplers has no advantages over grab water
sampling and other relatively low volume sampling techniques.
36
5. Factors influencing the uptake of compounds in SPMDs from air
5.1.1 RS of PCBs in air
In unpublished work described in this thesis, the RS of a number of PCBs in
SPMDs from air were tested by comparing SPMD samplers with
conventionally used MeVols. Instead of following the European standard
conditions the experimental setup to determine the RS was simplified for the
reasons mentioned above. Thus, the RS were determined in field under indoor
air conditions using only four SPMD samplers (instead of 24) and two
MeVols (instead of six) with pump rates of 7 and 14 m3 ⋅ d-1, respectively,
collecting the gas and particular phase separately. The indoor air sampling was
performed in a PCB-contaminated part of a laboratory building that were
closed for public use, where both the air concentration of PCBs and the
ambient temperature were considered to be quite constant. It is important
that the temperature is constant during the calibration since this variable can
affect the amount taken up in SPMDs in complex ways. At intermediate and
warm temperatures, the RS of compounds in SPMDs from water will increase
with increasing temperature (Booij et al., 2003; Huckins et al., 1999), while
Ockenden et al. (1998a) suggested that the uptake in SPMDs from air are
higher at cold compared to intermediate/warm temperatures. The
temperature also affects the partitioning between the gas-phase and particles,
and thus the amounts taken up in the SPMDs. Increasing temperature
increases vaporization and thus the partitioning into the gas-phase, while
decreasing temperature increases condensation onto particles which increases
the adsorption of compounds like PAHs to particles. Furthermore, the
ambient air concentrations vary differently with temperature depending on
the source of contamination. For instance, PAH-emissions from traffic are
highest during daytime, while emissions from wood combustions probably
increasing during the night.
In this calibration test the average air temperature was 22 °C with a standard
deviation of 2.3. The average amounts in standard 1-mL SPMDs (n=4), the
gas phase concentrations measured with MeVols and the exposure time (11
days) were used in eq. 10 to calculate the RS in m3 ⋅ d-1 for each compound.
The RS of the sum of each congener group (di- to octa) for standard 1-mL
SPMDs ranged from 2.5 to 4.5 m3 ⋅ d-1 (Figure 12). Thus, the RS of the PCB
congener groups were rather similar, as expected since all PCB compounds
(log KOW > 5) were probably under boundary layer control. These results are
in agreement with those of Shoeib and Harner (2002) who reported indoorcalibrated RS of PCBs, at the same temperature, to be usually in the range 3 to
5, and differed somewhat from outdoor-calibrated RS of PCBs, at an average
37
5. Factors influencing the uptake of compounds in SPMDs from air
temperature of 18 °C (range 7-31 °C) reported by Ockenden et al. (1998a).
Our results are also in agreement with Bartkow et al. (2004) who found
similar RS of 12 EPA PAHs (fluorene to benzo[g,h,i]perylene) in air (range 0.6
to 6.1 m3 ⋅ d-1). However, the concentration capacity (the KSA) of the SPMDs,
which was not tested in this work, will vary between each congener group.
Shoeib and Harner (2002) reported that equilibrium of PCB 28 (tri-CBs) and
PCB 52 (tetra-CB) was reached after 50-75 and 200-400 days, respectively,
while the uptake of PCBs with KOA values > 9 like PCBs 101 (penta-CB), 137
and 138 (hexa-CBs) was linear throughout the entire 450-day sampling period.
These results show that all PCB congeners tested in the work underlying this
thesis, were in the linear uptake phase during the entire 11-day calibration.
8,0
3
-1
RS (m d )
6,0
4,0
2,0
0,0
DiCBs
TriCBs
TetraCBs
PentaCBs
HexaCBs HeptaCBs
OctaCBs
Figure 12. The RS (m3 ⋅ d-1) of the sum of each PCB congener group at 22 °C
calculated from the average amounts (ng) in standard 1-mL SPMDs (n=4), and the
gas phase concentrations (ng ⋅ m-3) measured with active medium volume sampler
(MeVols) during a 11 days indoor air sampling.
5.2 Environmental variables
According to European standards for diffusive samplers of gases and vapours
the effect of the environmental variables relative humidity, temperature and
air velocity on the sampler performance (uptake rate) should all be tested
(Anonymous, 2002b). Temperature and wind-speed/turbulence should have
an effect on the uptake in SPMDs from air. In addition, chemicals bound to
particles or aerosols can be trapped on the membrane surfaces and affect the
amount taken up in the SPMD, and the UV light can decrease the amount
found in the SPMDs via photodegradation in air. The effect of environmental
variables on the uptake of compounds in SPMDs under ambient air
38
5. Factors influencing the uptake of compounds in SPMDs from air
conditions (excluding the effects of these factors under indoor air conditions)
are discussed in more detail below.
5.2.1 Wind-speed/turbulence
As previously mentioned, compounds with high log KOW values (KOW values
> approximately 4.4) have high KSPMD values and, thus, rapidly partition to the
membrane, and research have indicate that the uptake is restricted by a
boundary or diffusion layer at the membrane-water interface (Booij et al.,
1998; Huckins et al., 1999; Vrana and Schuurmann, 2002). Thus, the thickness
of the boundary layer and, consequently, the water flow/turbulence, affect the
uptake of these compounds, as were reported by Booij et al. (1998) and Vrana
and Schuurmann (2002). Vrana and Schuurmann (2002), for instance, found
that at low flow rates (0.06 to 0.28 cm ⋅ s-1), the uptake of hydrophobic
compounds (log KOW > 4) increased with increasing flow rate, but at
intermediate rates (0.28-1.14 cm ⋅ s-1) no significant increases in uptake of the
compounds were observed. In addition, Booij et al. (1998) reported three
times higher uptake rates of PAHs and PCBs under conditions of high
compared to low water flow/turbulence.
Although the diffusive mass transfer is much faster in air than in water, the
boundary layer is probably thicker and thus, wind-speeds/turbulence should
have an effect on the uptake in air as well. Figure 13 illustrates how the windspeeds/turbulence exposure affects the thickness of the boundary layer, and
thus the uptake of PAS. At extremely low wind-speeds/turbulence, when the
whole air compartment is quite still, the diffusive mass transfer in the thick
boundary layer is highly restricted and the turbulent transfer in low. Under
these conditions, the mass transfer to the sampler is so slow that a depletion
of the air at the boundary layer-air interface occurs. At increasing windspeeds/turbulence the thickness of the boundary layer decreases, and the
turbulent transfer in the air compartment increases. Thus, the uptake in the
sampler will increase. Under intermediate wind conditions, the boundary layer
and the turbulent transfer are quite constant, resulting in approximately
constant uptake in the sampler. At extremely high wind-speeds, as were tested
in the work underlying this thesis, the uptake increases with increasing windspeeds/turbulence since the turbulent transfer is high, and the thickness of
the boundary layer decreases significantly, which reduces the restriction to the
diffusive mass transfer in the boundary layer, so for SPMDs the relative
degree of control imposed by the membrane should progressively increase. In
39
5. Factors influencing the uptake of compounds in SPMDs from air
Uptake in the sampler (ng)
the most extreme wind-speeds/turbulence exposure cases, the boundary layer
can be eliminated and the uptake in SPMDs will then be solely controlled by
the membrane. Thus, the resistance to mass transfer in the boundary layer can
be tested by comparing the uptake in samplers at low and high air flows with
the uptake at intermediate wind-speeds/turbulence. In studies reported in
Papers II and III, the hypothesis that high wind-speeds (range 6-50 m ⋅ s-1)
increase the uptakes of PAHs and PCBs in SPMDs from air was tested.
Increase
(thick
boundary
layer)
Constant
(constant
boundary layer)
Increase
(thin
boundary
layer)
Wind-speed/turbulence exposure (m ⋅ s-1)
Figure 13. Correlation between wind-speeds/turbulence exposure, the thickness of
the boundary layer, and uptake in the PAS (modified from Anonymous, 2002b).
5.2.1.1 The wind effect test method
Considering the aforementioned difficulties, we concluded that the European
standard conditions for testing the wind effects on SPMD sampling of
nonpolar aromatic compounds in air could not be met. Instead, we tested the
wind effect in field, where the air volume capacity will be unlimited, and since
no point sources of PAHs and PCBs were present at the site, the gas
concentrations were reasonably constant between sampling points and over
time. More specifically, in the wind effect tests four (Paper II) or five (Paper
III) SPMD samplers were deployed simultaneously at different positions
inside a wind tunnel (Figure 14), housed in a large dark, un-heated building.
Each SPMD used was spiked with PRCs, as described in Papers II and III.
Two additional samplers, one deployed outside the wind tunnel (C1) and one
outside the building (C2), were used as controls of the uptake in SPMDs from
the almost still air inside the building and the uptake at ambient air
conditions, respectively. The wind tunnel was built with two compartments to
produce different wind-speeds in different parts of the tunnel. An electrical
40
5. Factors influencing the uptake of compounds in SPMDs from air
fan, mounted in the outside wall of the building at the inlet of the wind
tunnel, produced constant high wind flow during each sampling period. The
wind-speeds could not be measured at specific sampling sites due to high
turbulence in the wind tunnel, but the wind-speeds in the inlet (1400 × 1300
mm) and the outlet (400 × 500 mm) of the wind tunnel were calculated to be
6 and 50 m ⋅ s-1, respectively.
*
Figure 14. Experimental setup of the works described in Papers II and III, including
the two samplers used as controls of the uptake in SPMDs outside the wind tunnel
(C1) and outside the building (C2), the wind tunnel with two compartments (1300 ×
7000 mm and 500 × 2500 mm, respectively), the electrical fan (800 mm i.d.) mounted
in the outside wall of the building by the inlet of the wind tunnel*, and four (Paper
II) or five (Paper III) SPMD samplers inside the wind tunnel. One SPMD sampler
was deployed in the smaller (500 × 2500 mm) compartment of the wind-tunnel, and
the remaining three (Paper II) or four (Paper III) were distributed at equal distance
in the larger compartment of the wind-tunnel.
Since the tests were performed at ambient air conditions, the air temperatures
varied during each sampling period. However, measurements of the air
temperatures showed that they varied similarly between sites, and at each
sampling time, with average temperatures (measured in the unheated building,
in the wind tunnel and outdoors) being 15 °C in Paper II and 9 °C in Paper
III. Thus, the samplers in each study were exposed to very similar
temperature variations, and the air temperature should not have influenced
the differences in uptake between SPMDs. In conclusion, the use of this
method, which combined laboratory and field test conditions, fulfilled many
of the requirements of European standards (Anonymous, 2002b) although
the gas concentrations were not verified with any reference sampling methods
41
5. Factors influencing the uptake of compounds in SPMDs from air
or an independently calibrated measurement, and the wind effect was only
qualitatively determined.
5.2.1.2 Wind effect at high wind-speeds
After the 18-day (Paper II) or 21-day (Paper III) sampling to test the wind
effect on the uptake in SPMDs, the samples were analyzed for the PAHs and
PCBs listed in Papers II and III, respectively. The total amount of PAHs
taken up in the SPMDs in the wind tunnel were about two times higher than
the amount found in the SPMDs outside the wind tunnel (Figure 15). Similar
differences were observed for the uptake of PCBs, with about two to three
times higher total amounts of PCBs found in the SPMDs placed inside than
those outside the wind tunnel (Figure 15). There were differences in the
uptake of PAHs and PCBs inside the wind tunnel as well, with highest
uptakes observed at sites one (where turbulence was probably highest) and
four (Paper II)/five (Paper III) (where the calculated wind-speed was
highest). Highest uptakes of PAHs and PCBs were therefore observed in
those SPMDs exposed to the highest (assumed) turbulence/wind-speeds,
supporting the hypothesis that increasing wind-speeds increase the uptake in
SPMDs from air. Thus, the work described in this thesis showed that, in air,
the uptake of most nonpolar aromatic compounds (represented by the
studied PAHs and PCBs) in SPMDs is controlled by the boundary layer at the
membrane-air interface. However, a wind effect on the uptake in SPMDs was
not observed for all target compounds in the work of Papers II and III. The
uptakes of acenaphthene and fluorene, which had the lowest log KOA values
of the compounds studied, were similar both inside and outside the windtunnel, while the uptake of phenanthrene, with a log KOA value of 7.9 (Table
2, pp 55) (log KOW value of 4.5 (Mackay, 1992)) was higher in SPMDs inside
than outside the wind-tunnel. These results demonstrate that, in air, the
uptake of compounds with log KOA > 7.9 is controlled by the boundary layer.
To our knowledge, no wind effect on the uptake in SPMDs from air, and thus
no boundary layer control of the uptake of nonpolar air pollutants in SPMDs,
had been demonstrated prior to this work. However, Harner et al. (2003)
found that high wind increased the uptake of PCBs in unsheltered POGs. For
instance, the time to reach equilibrium for PCB 28 was five times higher in
still conditions than with an air flow of 4 m ⋅ s-1 in the cited study.
42
5. Factors influencing the uptake of compounds in SPMDs from air
1
PCB amount (ng SPMD -1 )
PAH amount (ng SPMD -1 )
1
d
0,8
0,6
0,4
0,2
c
a
0
a)
d
0,8
0,6
0,4
0,2
c
a,b
0
C1 C2 1
2
3
4
b)
5
C1 C2 1
2
3
4
5
Figure 15. Total amounts (ng ⋅ SPMD-1) of a) PAHs and b) PCBs taken up in the
replicate SPMDs in Paper III (first and second bars) and mean values of the amounts
taken up in the two SPMDs in Paper II (third bar). aThe first SPMD of C1 was
excluded due to sample losses during the GPC-analysis and bthe second SPMD was
excluded due to analytical interferences (these results can be found in Paper III). cIn
Paper II only four samplers were used. The fourth sampler in Paper II was deployed
at the same place as the fifth sampler in Paper III and thus, these data are compared.
dLow SPMD amounts probably due to damaged membrane.
The results reported in Paper II, showed that the differences in uptake
between the SPMDs inside and outside the wind tunnel increased with
increasing log KOA of PAHs up to 8.74. A possible explanation for these
results is that as hydrophobicity (and thus log KOA values) increases, the
restriction to diffusive mass transfer in the often aqueous boundary layer
(condensation at surfaces) increases, and thus high wind-speed in the wind
tunnel (which considerably reduces the thickness of the boundary layer) had a
greater effect on the uptake of more hydrophobic compounds. These
correlations were not observed for the PAHs with higher log KOA values,
which are mainly bound to particles, or for the PCBs. These results are not in
agreement with those found by Ockenden et al. (2001) who reported that the
wind effect on the uptake of PCBs (demonstrated by comparing the uptake in
protected SPMD samplers and those exposed to wind) decreased with
increasing chlorination (from tri- to octa-CBs). Their results indicate that the
uptake of more hydrophobic and larger PCBs is less affected by changes in
the thickness of the boundary layer, but this seems unlikely.
The SPMDs in the work presented in Papers II and III were spiked with
three 2H-PAHs and four 2H-PAHs together with four 13C-PCBs, respectively,
which were used to control the release at different wind-speeds. In all SPMD
43
5. Factors influencing the uptake of compounds in SPMDs from air
samples the levels of the PRC 2H-acenaphthene were below detection limit (<
3 × noise level). Thus, further discussion will only concern Paper III due to
the limited amount of PRC data in Paper II. In Paper III, the releases of the
remaining seven PRCs from the SPMDs inside the wind tunnel were up to
four times higher compared to the release from the SPMDs outside the wind
tunnel, and the highest amount were released from the SPMDs at site five,
where the calculated wind-speeds was highest. Highest loss rates were
therefore observed from the SPMDs exposed to the highest calculated windspeeds, demonstrating that the releases of PRCs were also restricted by the
boundary layer at the membrane-air interface.
5.2.1.3 Wind effect in field
A wind effect on the uptake and release in sheltered SPMDs, exposed to
extremely high wind-speeds/turbulence (6 to 50 m ⋅ s-1), was found (Papers II
and III). However, the wind-speeds/turbulence in the field will be more
intermediate, and less variable between locations. For example, the monthly
average ambient wind-speeds in Europe generally range between 1 and 10 m ⋅
s-1 (Anonymous, 2003). Thus, the wind effect in the field will be smaller. For
instance, in a field test by Ockenden et al. (2001) only small differences were
found when the uptake of PCBs in wind-sheltered SPMD samplers and
SPMD samplers exposed to wind were compared. Although the wind effect
on the SPMD sampling in the field will be generally low, extremely high windspeeds/turbulence can occur during sampling and, thus, the use of PRCs to
predict the site effect of wind is important.
5.2.2 Temperature
In water, Booij et al. (2003) have reported that the average RS of PCBs,
chlorobenzenes and PAHs were about three times higher at 30 °C than at 2
°C. Booij et al. (2003) also found a non-significant effect on the KSPMD at
these temperature differences, whereas the LPDE-water partition coefficients
were two times higher at 2 °C than at 30 °C. Huckins et al. (2002a) found that
the temperature effect on the KSPMD appear to be molecular structure specific
with a 2.6 fold differences in the KSPMD of phenanthrene measured at 8 and 30
°C, whereas no temperature effect was found on the measured KSPMD values
of PCB 52 and p,p´-DDE. In addition, Huckins et al. (1999) reported a 1.5fold increase in the RS of PAHs when the water temperature increased from
44
5. Factors influencing the uptake of compounds in SPMDs from air
10 to 26 °C. Thus, the temperature has a limited effect on the RS of
compounds in SPMDs from water.
Few studies have tested the temperature effect on the uptake in SPMDs from
air. Ockenden et al. (1998a) found that field-calibrated RS of PCBs were about
four times higher in winter than in summer. In winter, the ambient
temperatures ranged between -6 to 18 °C (mean 4 °C), while the temperature
range in summer was 7-31 °C (mean 18 °C). This temperature effect has
obviously never been observed on SPMD sampling in water, since this
compartment will be frozen at water temperature below ∼ 0 °C. Ockenden et
al. (1998a) suggested that the reason for these results was that at cold winter
temperatures decreased the solubility in air more than the solubility in the
lipid (triolein), and thus the lipid-air partition coefficient will increase. This
hypothesis was supported by Harner et al. (2003) who suggested that the
passive sampling medium-air partition coefficient (KPSM-A) will increase by a
factor of 2.5 to 3 for every 10 °C decrease in temperature. However, an
increase in the KSPMD will only increase the RS of compounds under
membrane-control, whereas for compounds under boundary layer control,
only the capacity of the SPMDs will increase with increasing KSPMD. Thus, the
effect of cold temperatures is rather complex and further research is needed
to understand the processes involved in the effect. It is clear though, that the
cold temperature effect can have as strong influence on summer and winter
data comparisons that are uncorrected for the temperature effect. For
instance, the highest and lowest temperatures measured in Umeå, Sweden
between 1996 and 2004 were 31.1 and -33.0 °C, respectively (Johansson,
2004). Research is also needed to assess the effects of intermediate and warm
ambient air temperatures. However, as previously mentioned, the effect on
the RS at intermediate and warm water temperatures is probably limited,
unless data on large geographical scales are compared, and we assume that air
temperatures above 0 °C affect the RS of compounds in air in a similar way as
in water.
5.2.3 Particles
In the work underlying this thesis, the amounts found the SPMDs were
expected to originate from air pollutants in the gas phase. However,
compounds bound to particles can contribute to the amounts found in the
SPMDs in several ways:
45
5. Factors influencing the uptake of compounds in SPMDs from air
I. Direct particle deposition on the exterior membrane surface during
sampling can allow compounds bound to particles to desorbs and
permeate through the membrane (particle-membrane transfer), and
thus increase the amounts found in the SPMDs.
II. Compounds bound to the particles at the exterior SPMD surface that
were not accumulated in the SPMDs during sampling, can be
extracted from the particles during dialysis instead and thus increase
the amounts found in the SPMD samples. The degree of sorption to
particles can vary for compounds with similar log KOA values because
of differences in their molecular structures. For instance, due to their
planar molecular structure PAHs can probably interact with soot
particles to a higher degree than PCBs and chlorobenzenes, which
have a more spherical structure.
In addition, particles can give poor dialytical recoveries of the target analytes
due to sorption to organic carbon (e.g., soot) at the exterior of the membrane
surface (Huckins et al., 1999). Thus, the cleaning of the exterior membranesurface prior to dialytical extraction is highly important since it reduces the
uncertainties of the SPMD data. This procedure will not reduce the effect of
particle-associated compounds that accumulate in the SPMDs during
sampling. However, when the SPMDs are protected by shelters the amounts
of particle-associated compounds taken up in the SPMDs should be limited.
For instance, Ockenden et al. (1998a) found that in comparison with HiVols,
more than 96 percent of the PCBs sampled by the SPMDs were associated
with the gas phase.
The potential of SPMDs to sample particle-associated compounds has been
investigated by not cleaning the exterior surface of the SPMD prior to
dialytical extraction (Bartkow et al., 2004; Lohmann et al., 2001). Bartkow et
al. (2004) found a good agreement between the air concentrations predicted
from the SPMD data and those measured with HiVols for the PAHs that are
mostly bound to particles. Lohmann et al. (2001) also detected PCDD/Fs and
PAHs in the SPMDs that are typically at least 85 percent associated with
particles, but predominantly gas phase compounds were found in the
samples. The reason for these results, i.e. whether the particle-associated
compounds in the samples originated from actual SPMD uptake of desorbed
compounds or from dialytical extraction, was not clear, and further research is
needed to understand the sampling processes of particle-associated
compounds.
46
5. Factors influencing the uptake of compounds in SPMDs from air
5.2.4 UV sunlight
A fourth environmental variable that can affect the amount of analytes found
in the SPMD air samples is UV sunlight. Field tests by Orazio et al. (2002)
demonstrated that in SPMDs exposed to direct sunlight, some of the 16 EPA
PAHs were photolyzed in just 2 minutes, while in SPMDs sheltered inside
foil-topped canisters, for a one-week air sampling, the PAHs degraded
severely. In addition, Bartkow et al. (2004) found higher losses of the 2HPAHs anthracene and pyrene than expected, probably due to
photodegradation during sampling, although the SPMDs (820 × 28 mm
LDPE tubes filled with 1 mL of triolein) were covered by chambers. Thus,
photodegradation of the compounds accumulated in the SPMDs may occur
during both sampling and retrieval. It is therefore critical to ensure that the
SPMDs are not exposed to any direct or reflected sunlight, and that the
sampler design used reduces the UV sunlight sufficiently. In addition, the site
effects of UV sunlight can vary significantly, both temporally and
geographically. In Umeå, 2004, for instance, the UV radiation (W ⋅ m-2) on the
18th of January was measurable only between 10 am and 4 pm with a
maximum below 75 W ⋅ m-2, while the UV radiation in the 15th of July was
measurable from 1 am to 10 pm with maximum value around 750 W ⋅ m-2
(values above 300 W ⋅ m-2 mean that the sun is shining) (Johansson, 2004). It
is therefore also important to choose sampling locations that are in the shade
as much as possible.
47
48
6. Approaches to reduce the site effects of environmental variables
6. APPROACHES TO REDUCE THE SITE EFFECTS OF
ENVIRONMENTAL VARIABLES
As previously discussed, several environmental variables can affect the uptake
in SPMDs during in the field deployments and thus cause deviations from
laboratory pre-calibrated RS of the target analytes. In order to compare
temporally or geographically different samples it is important to reduce and
assess the site effects of these environmental variables.
6.1 The use of shelter
In the work underlying this thesis, the Exposmeter air sampler, described in
detail previously, was used (Figure 16). Included in the design of the sampler
was a metal umbrella that was used to protect the SPMDs from direct
exposure to wind, UV light, particle deposition and rain, but allowing air to
pass under and around the SPMDs. Examples of other type of shelter designs
that have been used to protect the SPMDs from direct exposure are
aluminium boxes (dimensions 300 × 300 × 300 mm) called Stevenson screens
(Ockenden et al., 2001) and metal boxes that measuring 110 (h) × 220 (l) × 65
(d) mm (Lohmann et al., 2001).
Figure 16. Design of the metal umbrella (320/160 mm i.d. × 220 mm h) used to
shelter the SPMDs in the work underlying this thesis.
The work presented in this thesis tested whether the metal umbrella, designed
to reduce the wind-speeds/turbulence inside the sampler, decreased the
uptake and release rates in SPMDs at high wind-speeds/turbulence (Paper
III). The hypothesis was tested by comparing the uptake and release in
49
6. Approaches to reduce the site effects of environmental variables
unsheltered SPMDs and SPMDs sheltered by metal umbrellas at high windspeeds/turbulence. In all, five samplers, each including two SPMDs inside a
metal umbrella and one SPMD on top, outside the umbrella, were deployed
simultaneously for 21 days in a wind tunnel with high air flow (Figure 14).
6.1.1 Reduction of the wind effect by shelter
Comparison between unsheltered SPMDs and SPMDs sheltered by metal
umbrellas showed that the use of this shelter design reduced the total uptake
of PAHs and PCBs by 38 and 55 percent on average, respectively (Figure 17).
Furthermore, the release of 13C-PCB 15 and 2H-phenanthrene decreased by 025 and 13-22 percent, respectively, when the SPMDs were protected by the
metal umbrella. Thus, the uptake and release in the SPMDs decreased when
they were sheltered by metal umbrellas, demonstrating that the use of the
metal umbrella reduced the effect of high wind-speeds/turbulence.
Furthermore, the difference between the lowest and highest amounts of
PAHs and PCBs taken up in the SPMDs was reduced by 10 and 8 percent,
respectively, when the SPMDs were covered with the metal umbrella. These
results show that the metal umbrella also reduced the variability in wind effect
between sites. However, this test was performed under extremely high and
variable wind-speeds/turbulence. Under normal wind conditions, the rate and
variation in wind inside the metal umbrella will be much lower, and thus the
variability in uptake between sites will be even lower. These results are in
agreement with Ockenden et al. (2001), who reported wind-speeds inside a
so-called Stevenson’s screen that were below the detection limit (< 0.4 m · s1
), while the wind-speeds outside the devices were 2.6 to 31.3 m · s-1. In
addition, Harner et al. (2003) found that the effect of high wind was reduced
to insignificant when POGs were sheltered by chambers.
50
6. Approaches to reduce the site effects of environmental variables
125
PCB amount (ng SPMD -1 )
PAH amount (ng SPMD -1 )
2500
2000
1500
1000
500
0
a)
1
2
3
4
5
b)
100
75
50
25
0
1
2
3
4
5
Figure 17. Total amounts (ng ⋅ SPMD-1) of a) PAHs and b) PCBs in SPMDs
sheltered by metal umbrellas (average of two SPMD measurements) (left bars) and in
unsheltered SPMDs (right bars) at high wind-speeds/turbulence (modified from
Paper III).
The metal umbrella reduced the uptake of the native compounds more than
corresponding releases of the PRCs. There are two possible reasons for these
results and most likely both of them contribute to the differences in uptake
and release between sheltered and unsheltered SPMDs. Firstly, the
unprotected SPMDs may have been exposed to such high wind-speeds that
the boundary layer at the membrane-air interfaces was very thin or nonexistent. Under these sampling conditions the mass transfer of the native
compounds in the bulk atmosphere was probably not restricted by the
boundary layer and the native compounds were actively transferred to the
membrane surface by the high air flow/turbulence. In contrast, the PRCs in
the bulk lipid of the SPMD were transferred to the membrane surface by
molecular diffusion. Since active transport is a faster transfer process than
molecular diffusion, the amount of PRCs transferred to the membrane-lipid
interface was probably lower than the amount of native compounds
transferred to the membrane-air interface, and thus the PRC release rates
could not fully reflect the increasing uptake rates observed in the unprotected
SPMDs. Secondly, the unsheltered SPMDs may have been affected by
particle deposition on the membrane surface to a much higher degree than
the sheltered SPMDs, because of their direct exposure to high air flows.
Under such conditions, the amount taken up in the unprotected SPMDs is
not only affected by the high wind, but also by direct particle-membrane
transfer of compounds: a process that cannot be reflected by the release of
PRCs. This hypothesis was supported by the results since the contribution of
51
6. Approaches to reduce the site effects of environmental variables
the 4- and 5-ring PAHs to the total uptake of the PAHs increased by 20
percent in unsheltered SPMDs. The effect of increasing direct particle
deposition on the uptake of PCBs should be limited since these compounds
are predominantly found in the gas phase. However, further research is
needed to verify the anisotropy observed between release and uptake rates of
unprotected SPMDs exposed to high wind, and to define the uptake and
release mechanisms at these sampling conditions. In conclusion, the results of
this study demonstrate that the uptake in unsheltered SPMDs was affected by
the high wind, and probably also by particle-deposition, and that these effects
were reduced by using the metal umbrella. However, the sampler design used
in the work underlying this thesis is probably not optimal for reducing the
effect of UV sunlight since the shelter consists of metal, which will reflect the
sunlight inside the devices. One way to improve the design of the metal
umbrella would be to cover the inside of the metal umbrella with a dark
material.
6.2 The use of performance reference compounds (PRCs)
In water, several studies have indicated that PRCs can be used to assess the
site effect of water flow/turbulence, temperature and biofouling (Huckins et
al., 1993; Booij et al., 1998; Huckins et al., 2002a; 2002b; Vrana and
Schuurmann, 2002; Booij et al., 2003). The use of PRCs to assess the site
effect of wind-speed/turbulence on the uptake in SPMDs from air will be
discussed below. The possibility of using PRCs to assess the effect of
particles, temperature and photolysis will also be briefly discussed.
6.2.1 PRCs´ utility for assessing wind effect
The work described in Papers II and III tested if the releases of PRCs were
related to the uptake of target analytes under variable wind conditions (from
still air to an air flow of 50 m ⋅ s-1). The native target compounds were a
number of PAHs and PCBs, while the 2H-PAHs acenaphthene, phenanthrene
and pyrene (Paper II), and the 2H-PAHs acenaphthene, fluorene,
phenanthrene and pyrene together with the 13C-PCBs 3, 15, 37 and 54 (Paper
III), respectively, were used as PRCs. These labelled compounds were
selected in order to test the use of PRCs covering as a wide range of vapour
pressures as possible with releases that were suitable for different detectors
and for sampling periods of varying duration, e.g., one week to several
52
6. Approaches to reduce the site effects of environmental variables
months. After 18 days (Paper II) or 21 days (Paper III) sampling, the levels of
the PRC 2H-acenaphthene was below its detection limit (< 3 × noise level) in
all SPMDs. In Paper II, the precision of the PRC 2H-pyrene data was low due
to analytical interferences in the HRGC-LRMS analysis. Thus, further
discussion will only concern Paper III due to the limited amount PRC data in
Paper II.
As mentioned previously, both the releases and the uptakes were highest by
the SPMDs exposed to the highest calculated wind-speeds, demonstrating
that both the uptake of the compounds, and the release of selected PRCs,
were controlled by the boundary layer. Thus, the release rates of PRCs were
related to the uptake rates of native compounds, showing that they could be
used to assess the wind effect. These results are in agreement with Vrana and
Schuurmann (2002) who reported that in water, the release of 2H-anthracene
increased three-fold when the water flow increased five-fold from 0.06 to
0.28 cm ⋅ s-1. Also Booij et al. (1998) found higher releases of PCB 29, 2Hphenanthrene and 2H-chrysene at high water flow (30 cm ⋅ s-1) than under
conditions of low flow (0.03 cm ⋅ s-1). However, the releases of PRCs were
not able to fully assess the uptakes of compounds in unsheltered SPMDs
probably due to the high turbulent transfer of the native compounds and high
particle-deposition on the membrane surface that facilitate direct particlemembrane transfer (effects further discussed in 5.2.3). Thus, in order to use
PRCs to predict the effects of high wind it is critical that the SPMDs are
protected by metal umbrellas. In addition, both Papers II and III revealed
problems with the precision of the PRCs due to analytical interferences. Booij
et al. (1998) used PRCs and found that they allowed the uptake kinetics of
PAHs to be predicted, but with poor precision due to analytical interferences.
Thus, when using PRCs to calibrate the SPMD sampling in situ robust
analytical quality control must be applied.
The releases of PRCs generally increased with decreasing log KOA of their
analogous native compound (Figure 18). (Figure 18 also shows that the
amount of the PRC released for each compound (log KOA value) was affected
by the wind-speeds, see differences in the y-values in the figure). Hence, the
SPMD affinity, and thus the KSA, increased with the KOA values of the PRCs
used. More specifically, acenaphthene, which had the lowest KOA value, was
completely released before the sampling was ended, the releases of 2Hfluorene and 13C-PCB 3 were mostly > 80 percent, the releases of 2Hphenanthrene and 13C-PCB 15 were in most cases between 20 and 80 percent,
while the releases of the 2H-pyrene and 13C-PCBs 54 and 37 were within the
53
6. Approaches to reduce the site effects of environmental variables
20-80 percent range or below 20 percent. These results are consistent with the
findings of Ockenden et al. (2001) who observed losses of PCBs 28, 52, 101
and 153 that were 85, 50, 25 and 8 percent, respectively, and non-measurable
losses of 13C-PCBs 138 and 180, after a four-month sampling campaign, due
to their high KOA values. Consequently, acenaphthene or compounds with
similar KOA values should be used as PRCs in short-time (days) deployments,
for intermediate-time (weeks) deployments compounds like 2H-phenanthrene
and 13C-PCBs 15 and 28 should be used, for longer-time deployments
(months to years) compounds like 2H-pyrene and the 13C-PCBs 54 and 37
should be used, and for even longer sampling periods (years) compounds
with slightly higher log KOA values should be used.
Amount released (%)
100
80
60
40
20
0
6
7
8
9
10
log K OA
Figure 18. Average amount (%) of PRC released from the replicate SPMDs in the
wind tunnel (filled circles), outside the wind tunnel [C1, (triangles)] and outside the
building [C2 (squares)] as a function of the log octanol-air partition coefficient (log
KOA) of their analogous native compound (Table 2) (Paper III).
The ranges of the ke values calculated using eq. 12, and the SPMD data with
remaining PRC-amounts that were 20 to 80 percent of the initial PRCamounts, are presented in Table 2. The ke values of the 2H-PAHs were lower
than those calculated at 22 °C and wind-speeds between 0.8 and 1.7 m ⋅ s-1 for
the 2H-PAHs anthracene (0.090 d-1) and pyrene (0.094 d-1) by Bartkow et al.
(2004). The cited authors suggested that their loss rates were too high. In
addition, they found similar release rates of anthracene and pyrene, although
these compounds differ in their log KOA values. As previously mentioned,
they suggested that one possible reason for their results was photolysis of the
PRCs used during sampling. Ockenden et al. (2001) calculated the ke values
for PCBs 28, 52, 101 and 153, at 15 ºC and wind-speeds in the range 2.6-31.3
m ⋅ s-1, to be 0.0085, 0.0047, 0.002 and 0.0007 d-1, respectively. Thus, our
54
6. Approaches to reduce the site effects of environmental variables
calculated ke values were about an order of magnitude higher than theirs.
These results could be due to the wind-speeds being much higher during our
study. Another possible explanation is that sampling devices with different
designs were used, causing either the SPMDs in our study to be more directly
exposed to the high air flow, or the SPMDs in their study to be exposed to
very still air, and thus their loss rates to be too low.
Table 2. ke values of performance reference compounds (PRCs)
(excluding acenaphthene) estimated in Paper III, and the log KOA
values of their analogous native compounds.
log KOA
ke values (d-1)d
13C-PCBs
PCB 3
7.0a
0.036-0.061
PCB 15
7.9a
0.018-0.060
PCB 37
9.0b
0.011-0.033
PCB 54
7.3b
0.016-0.042
2H-PAHs
c
Fluorene
7.1
0.036-0.063
Phenanthrene
7.9c
0.012-0.065
Pyrene
9.2c
0.011-0.018
Log octanol-air partition coefficient (log KOA) at 20 °C , ameasured by
Harner and Bidleman (1996), cmeasured by Harner and Bidleman
(1998), and bpredicted by Chen et al. (2002), respectively. dExchange
rate constant at 9 °C and different wind-speeds.
Since we assumed that all SPMDs were exposed to the same air, the ability of
the PRCs used to compensate for the differences in wind-speeds in the windtunnel was tested by calculating the air concentrations of some selected PAHs
and PCBs using the in situ calibrated ke values of 2H-phenanthrene and 13CPCB 15, and eq. 13, and then comparing the variability of these values. There
were several factors to consider in these calculations. First, the releases of the
PRC 2H-phenanthrene were > 60 percent for the SPMDs inside the wind
tunnel. Thus, the uptake of phenanthrene was nonlinear, and the use of the
linear uptake model described in eq. 13 would cause the phenanthrene
concentrations calculated to be underestimated. Second, the KSA values of the
selected compounds were not available and their KOA values had to be used as
surrogates. The sources of the surrogate PAH and PCB KOA values were
Harner and Bidleman (1998), and Chen et al. (2002), respectively. Since the air
concentrations were only semi-quantitatively determined (by using KOA
values) data were normalized to the highest concentration determined for
each compound (Figure 19).
The CV for the calculated PAH- and PCB-concentrations ranged between 28
to 41 and 28 to 46 percent, respectively (Figure 19). Thus, the use of PRCs to
calibrate the ke values in situ and thereby correct for the site effect of wind
reduced the between-site differences to less than 50 percent from as much as
55
6. Approaches to reduce the site effects of environmental variables
two- and three-fold difference in the total amount of PCBs and PAHs,
respectively, found in the SPMDs. If the use of PRCs had fully compensated
for the effect of different wind-speeds, the quantified air concentrations
would have been equal between sites. However, we consider that the CV
(introduced by the 21-day sampling and the chemical analysis) of the
estimated air concentrations were good. Thus, the utility of the PRCs used for
assessing the site effect of wind was high.
Phenanthrene (CV: 28 %)
PCB 52 (CV: 28 %)
1,0
1,0
0,8
0,8
*
0,6
0,6
0,4
0,4
0,2
0,2
0,0
*
0,0
1
2
3
4
5
1
Pyrene (CV: 29 %)
2
3
4
5
PCB 80 (CV: 29 %)
1,0
1,0
0,8
0,8
*
0,6
0,6
0,4
0,4
0,2
0,2
0,0
*
0,0
1
2
3
4
5
1
Fluoranthene (CV: 41 %)
3
4
5
PCB 101 (CV: 46 %)
1,0
1,0
0,8
0,8
0,6
2
0,6
*
0,4
0,4
0,2
0,2
0,0
*
0,0
1
2
3
4
5
1
2
3
4
5
Figure 19. Normalized air concentrations of six PAHs and PCBs estimated using the
in situ calibrated ke values of 2H-phenanthrene and 13C- PCB 15, respectively, and the
KOA values of PAHs and PCBs suggested by Harner and Bidleman (1998) and Chen
et al. (2002), respectively, as surrogates for unavailable KSA values (Paper III). The
coefficient of variance (CV), derived from 21-day sampling and chemical analysis, of
the estimated air concentrations of each compound was shown in brackets. *Low
uptake, and high ke values, probably due to membrane damaged.
56
6. Approaches to reduce the site effects of environmental variables
6.2.2 PRCs´ utility for assessing site effects in air
The work described in this thesis demonstrated the high potential of the PRC
approach to assess the site effect of wind. In order to achieve high precision
of the PRCs, the SPMDs have to be sheltered from direct wind exposure and
robust analytical quality control should be applied. However, the reduced
correlation between the releases of PRCs and the uptakes of native
compounds in unsheltered SPMDs indicate that the effects of particles may
not be predicted by the use of PRCs (Paper III). To our knowledge, the use
of the PRC-approach to predict the effect of temperature and photolysis has
not been tested. It has been suggested, however, that photosensitive
compounds with high SPMD affinity should be added to the SPMD prior to
membrane enclosure, to ensure that compounds accumulated in the SPMD
do not photodegrade during sampling and retrieval (Bartkow et al., 2004).
However, this approach will only prove if photodegradation occurred or not,
rather than quantitatively determine its effects on the uptake of specific
compounds.
57
58
7. Fields of application and sampling strategies
7. FIELDS OF APPLICATION AND SAMPLING STATEGIES
SPMDs have been used in a number of different applications and
environments, but primarily in water (Table 1). The sampling strategy varies
depending on the purpose of the study. The fields of application for which
SPMDs have been used in the work described in this thesis are presented
below, and the chosen sampling strategies are discussed.
7.1 Determination of local/regional atmospheric distribution
The atmospheric distribution of the gas phase PAHs in the Bangkok region,
Thailand, was investigated in the work presented in Paper I. SPMD samplers
were deployed for three weeks at six different locations (A-F). The sites were
chosen in order to cover a wide range of pollution situations ranging from
remote areas with background air quality to urban areas. The SPMD samples
were analyzed for 14 of the 16 EPA PAHs (excluding naphthalene and
acenaphthylene) together with benzo[e]pyrene and 1-methylphenanthrene.
The total amounts of the PAHs (excluding 1-methylphenanthrene) found in
the SPMDs varied between 17 and 134 ng ⋅ SPMD-1 ⋅ d-1, and were correlated
with the assumed degree of pollution at the sampling sites (Figure 20). For
example, SPMDs at the urban sites E and F sequestered about ten times more
of the PAHs than those deployed at site A (rural area). The predominant
individual PAH was phenanthrene, for which the amounts in the SPMDs
ranged between 9.8 and 64 ng ⋅ SPMD-1 ⋅ d-1.
59
7. Fields of application and sampling strategies
-1
-1
PAH amounts (ng SPMD d )
160
120
80
40
0
A
B
C
D
E
F
Figure 20. Total amounts of 14 EPA PAHs (excluding naphthalene and
acenaphthylene) taken up in SPMDs from the atmosphere of the Bangkok region,
Thailand (pollution was assumed to increase from sites A to F) (modified from Paper
I).
In the work reported in Paper IV, the distribution of the gas phase
concentrations of PAHs and nitro-PAHs were monitored in five European
countries (Austria, the Czech Republic, Poland, Slovakia and Sweden). The
study is described in more detail below (see 7.2). Briefly, however, SPMD
samplers were deployed during two three-week sampling periods at eight
locations in each country. The SPMD data collected within each country
differed 5- to 20-fold in PAH-levels, and 6- to 30-fold in nitro-PAH-levels.
The highest levels were detected at sites defined as urban, while the lowest
levels were detected at remote/rural sites, showing that the variation in the
PAH- and nitro-PAH levels within each country followed the subjectivelydefined pollution situations.
When using SPMDs to determine the distribution of gas phase
concentrations at a local scale, the site effects of environmental variables are
generally low, and the variations in the SPMD data reflect the spatial
distribution. For example, the average sampling temperatures reported in
Paper I was in the range 29.1 to 31.0 °C. However, when comparing winter
and summer data, as mentioned above, the temperature effect on the uptake
in SPMDs should be considered. For instance, the temporal differences in
Poland were high, with uptake of PAH- and nitro-PAHs being one and two
orders of magnitude higher, respectively, in winter (average 0.7 °C) than in
summer (average 16 °C) (Paper IV). One possible reason for these results is
60
7. Fields of application and sampling strategies
that the RS of PAHs were probably higher in the winter than in the summer.
However, Ockenden et al. (1998a) suggested that the RS of PCBs in summer
and winter differed only about four times and thus, the temperature effect
cannot be the main reason for the high temporal variations found in Poland.
Instead, the main reason was probably that the emissions from coal
combustion increased in the winter since residential heating was required
when the air temperatures were cold. The results of the work underlying this
thesis (Papers I and IV) demonstrate that SPMDs are efficient samplers of
gas phase pollutants, and are able to determine the spatial distribution of gas
phase PAHs in rather restricted areas. The high efficiency of SPMDs to
detect differences in atmospheric concentrations at a local scale has also been
reported by Lohmann et al. (2001). In addition, the use of the SPMD
technique, which does not require electricity, made sampling possible at
remote/rural areas where the infrastructure was limited, like the countryside
in Thailand. However, due to the lack of calibration data at the time, the
atmospheric concentrations could not be quantitatively determined. Instead,
the RS values of PAHs in water calibrated by Huckins et al. (1999) were used
in Paper I for preliminary calculation of the PAH concentrations in air.
7.2 Determination of continental atmospheric distribution
In the work of Paper IV, the gas phase PAHs and nitro-PAHs in air of five
European countries (Austria, the Czech Republic, Poland, Slovakia and
Sweden) were measured during two sampling campaigns in order to
determine whether there were spatial differences between northern, central
and eastern Europe. Prior to the air sampling campaigns, the research groups
of each participating country received written sampling instructions and
sampling equipment from the Environmental Chemistry, Umeå University,
Umeå, Sweden. According to instructions, the SPMD samplers were deployed
by each research group for two three-week periods, one in the autumn of
1999 and the other in the summer of 2000, except in Poland where four-week
samples were taken (at different locations) in the winter of 1999 and the
summer of 2000. In all, SPMD samplers were deployed at 40 locations; eight
sites per country, at similar periods of time in two different seasons (Figure
21). The eight sampling sites were chosen by the research group from each
country following the guidance to cover a wide range of pollutant situations
from remote and rural sites to urban and industrial areas. The average
sampling temperature of each location was measured in parallel and other
environmental conditions such as wind-speed and wind direction were
measured, when possible (data listed in Table 1, Paper IV). The SPMD
61
7. Fields of application and sampling strategies
samples were analyzed for 13 of the 16 EPA PAHs (excluding
acenaphthylene, fluoranthene and pyrene) and the six nitro-PAHs 1nitronaphthalene (1-NN), 2-nitronaphthalene (2-NN), 2-nitrofluorene, 9nitroanthracene, 3-nitrofluoranthene, and 1-nitropyrene. However, in 2000,
all SPMDs were spiked with native fluorene, so this compound was not
analyzed in these samples.
N
W
N
E
W
S
E
S
S PL1-2
W PL1-2, 4-8
S PL3-6
Warsaw
W PL3
S PL7-8
SE4, 6
POLAND
SE2
CZ4
SE8
SE7
SE3
Prague
CZ6
CZ3
S
#
CZECH REPUBLIC
CZ7
SE1
CZ2
S CZ1
#
CZ5
SK1
CZ8
SK8
SLOVAKIA
SE5
AU5
AU7
AU4
Vienna
SWEDEN
SK6
SK3
SK2
Bratislava
SK4-5
SK7
AU8
AU3
AU2
AUSTRIA
AU1
AU6
Stockholm
0
100
200
300
400
500 Kilometers
0
100
200
300
400
500 Kilometers
Figure 21. Locations of the sampling sites in Sweden (SE), Austria (AU), the Czech
Republic (CZ), Slovakia (SK) and Poland (winter 1999, W-PL, and summer 2000, SPL) (Paper IV).
The total amounts of PAHs sequestered by the SPMDs in 1999 and 2000
ranged between 6.3-1200 and 5.0-1000 ng ⋅ SPMD-1 ⋅ d-1, respectively, and the
total amounts of nitro-PAHs range between 0.036-4.0 and 0.0011-2.7 ng ⋅
SPMD-1 ⋅ d-1, respectively. Phenanthrene was the most abundant individual
PAH found in the SPMDs, while the three nitro-PAHs 1-NN, 2-NN and 2nitrofluorene dominated the nitro-PAH-patterns. These SPMD air data
showed a wider range than those reported by Lohmann et al. (2001) who
measured gas phase concentration of PAHs at 11 sites in the northwest of
England. To our knowledge, passive sampling of nitro-PAHs using SPMDs
has not previously been reported. However, the levels detected were about a
tenth of the gas phase concentrations of 1-NN and 2-NN measured with an
62
7. Fields of application and sampling strategies
active high-volume sampler system in the two cities Birmingham, UK, and
Damascus, Syria (Dimashki et al., 2000).
The SPMDs deployed at 40 different locations from northern to central and
eastern Europe showed that the levels measured in East Europe (Poland,
Slovakia, and the Czech Republic) were about ten times higher than those
found in the sampled areas of northern Europe (Sweden), whereas the
measured levels of gas phase nitro-PAHs were more than ten times higher
than the levels measured in northern (Sweden) and central (Austria) Europe.
Thus, the spatial differences in the gas phase concentrations of PAHs and
nitro-PAHs between the five countries were high, and the continental
distribution of nitro-PAH and PAH-levels were similar. The spatial
distribution of gas phase PAHs in Europe had not been measured with
SPMDs prior to this study. However, the spatial variations in both PAH- and
nitro-PAH concentrations found were similar to the variations in atmospheric
PAH-concentrations measured across Europe by Jaward et al. (2004b) using
PUF disks.
The differences in the individual and total PAH-levels found in the SPMDs
deployed at the 40 locations were visualized in a score plot of the first two
PCs (explaining 74 and 9.6 percent, respectively, of the variance) of a PCA
model presented in the work of Paper IV (Figure 22). On the left-hand side
of the score plot, sampling sites in Sweden and Austria were predominantly
found, whereas sampling sites of the Czech Republic and Slovakia were found
on the right-hand side. The corresponding loading plot revealed that at the
sampling sites to the right in the score plot, viz. East Europe (the Czech
Republic (CZ), Slovakia (SK) and Poland 1999 (PL-W)), the levels of e.g.,
phenanthrene, chrysene and the total PAH-s found in the SPMDs were
higher than those found in Sweden, Austria and Poland 2000 (PL-S).
63
7. Fields of application and sampling strategies
AU6-00
AU2-00
SE2-00
AU3-00
2
SK6-00 CZ5-00
PL5-S
CZ1-99
PL7-S
SK7-00
SK3-99
AU4-00
PL3-S
CZ6-99
SK7-99
SE5-00
PL6-S
PL4-S
CZ7-00 CZ2-99
SK3-00
SE4-00
AU3-99
SK2-99
PL2-S
PL1-S
SE1-00 PL3-W
SK8-00
SE3-00 SK8-99
SK1-00
SK6-99
SE6-00
AU1-00
SK5-99
PL2-W
AU7-00
AU5-00
CZ6-00
CZ2-00
CZ1-00
PL8-S AU4-99 CZ8-00
PL6-W
SE7-00
AU8-00
CZ4-00
CZ3-99
SK4-00
PL4-W
SK1-99
CZ4-99
CZ5-99
CZ8-99
SK4-99
SK5-00
PL5-W
CZ7-99
PL7-W
SK2-00
AU6-99
PL8-W
SE8-00
AU7-99
AU8-99
SE1-99
PL1-W
SE5-99
SE3-99
SE4-99
AU1-99
CZ3-00
AU5-99
SE2-99
SE6-99
SE7-99
SE8-99
PC2
1
0
-1
-2
-7
-6
-5
-4
-3
-2
-1
0
PC1
1
2
3
4
5
6
7
Figure 22. Score plot of the first two principal components (PCs) of a PCA model
visualizing the correlation between the individual and total PAH-levels found in the
SPMDs deployed in Sweden (SW), Austria (AU), Poland (PL-W and PL-S), Slovakia
(SK) and the Czech Republic (CZ) during the 1999 and 2000 sampling campaigns
(Paper IV). The corresponding loading plot shows that at the sampling sites to the
right in the score plot the levels of e.g., phenanthrene, chrysene and the total
amounts of PAHs were higher than those found at the sites to the left in the score
plot.
The measured weather data reported in Paper IV (Table 1, Paper IV) showed
that for each sampling campaign the environmental sampling conditions like
air temperatures were generally similar between the sampling locations. Thus,
the site effect on the uptake in the SPMDs was limited, and the variations in
the SPMD data reflected the spatial variations between the five European
countries. These results demonstrate that SPMDs are suitable for integrative
sampling of the atmospheric distribution of nonpolar air compounds like
PAHs at a continental scale. Results of Paper IV also demonstrates the
advantage of using SPMDs in monitoring programs compared to
conventionally used HiVols. The SPMD´s low cost, and no need of special
trained personal or maintenance during sampling made it possible to
accomplish a three-week integrative sampling at 40 locations during two
sampling campaigns collecting a total of 80 samples. Such numbers of
samples can give a much more complete picture of the spatial distribution of
POPs in air across Europe than, for instance, the eight sites distributed in six
countries that were included in the air monitoring of POPs in air and aerosol
co-ordinated by CCC/EMEP in 2002 (Aas and Breivik, 2004). The utility of
SPMDs to measure the continental distribution has been shown in two other
studies by Ockenden et al. (1998c) and Meijer et al. (2003) where SPMDs
were deployed at 10 and 11 remote sites from the south of the UK to the
64
7. Fields of application and sampling strategies
north of Norway for two two-year sampling campaigns in the periods 1994 1996 and 1998 - 2000, respectively, in which PCBs and chlorinated pesticides
were measured. However, to date, most SPMD air data have been interpreted
by comparing the differences in amounts and profiles found in the SPMDs.
To further improve the utility of SPMDs it is important that PRCs are used in
the sampling protocols, and calibration data are further developed so that the
spatial distribution at both local and global scales can be quantitatively
determined.
7.3 Determination of pollution sources
SPMDs have been used to determine the source of pollution in several
studies (Table 1). Different approaches to determine the source of the PAHs
in air measured by SPMDs are discussed below.
7.3.1 The use of individual compound ratios
One of the approaches used to determine the source of pollution in the
atmosphere is to compare levels of a few individual compounds. In the work
underlying this thesis, phenanthrene, fluoranthene and pyrene, respectively,
were the most abundant individual PAH compounds found in the SPMDs.
The ratio of fluoranthene to pyrene has been suggested as a useful indicator
of traffic as the source to the atmospheric concentrations of PAHs measured
with HiVols (e.g., Yunker et al., 2002). However, when testing this approach
in Paper I, where the PAH-levels in air at different levels of traffic intensity
were sampled, the fluoranthene/pyrene ratios were not correlated with the
assumed degree of traffic intensity, probably because this approach was
developed for HiVols, which sample both gas and particle concentrations.
Other approaches that have been suggested to detect if traffic is the major
source of atmospheric PAH-levels are to use ratios of the particle associated
PAHs benzo[g,h,i]perylene to benzo[e]pyrene and coronene to
benzo[e]pyrene (Lim et al., 1999; Nielsen, 1996), or to use the benzo[e]pyrene
levels as indicators (Müller et al., 1998). Thus, many of the PAH-ratios used
today to detect if motor vehicles are the major PAH source are most suitable
for HiVols, which sample the total atmospheric concentration (both the gas
and the particle phase).
65
7. Fields of application and sampling strategies
PAS predominantly sample compounds in the gas phase and, due to the size
limitation in the membrane of the SPMDs, only the primary bioavailable
compounds can accumulate in the SPMDs. Thus, new indicative individual
compound ratios have to be found that are suitable for SPMD sampling.
When selecting suitable compound ratios for SPMD sampling it is important
to consider that the RS of compounds is affected by their physical/chemical
properties. Phenanthrene accounted for about 30 to 50 percent of the total
PAH amounts found in the SPMDs (Papers I-V). High concentrations of
alkyl-PAHs compared to unsubstituted PAHs have been suggested to indicate
the presence of petrogenic sources (e.g., Ramdahl, 1983; Rogge et al., 1993;
Nielsen, 1996; Lim et al., 1999, McDonald et al., 2000). Thus, since the RS
values of phenanthrene and methylphenanthrene are quite similar (4.46 vs
⋅
d-1
(Luellen
and
Shea,
2002),
the
5.14
L
water
methylphenanthrene/phenanthrene ratio is a possible indicator of the
contribution of traffic to PAH concentrations in air sampled with SPMDs. In
the
work
of
Paper
I
it
was
shown
that
the
1methylphenanthrene/phenanthrene ratio increased with increasing total PAH
amounts in the SPMDs, and assumed traffic intensities (Figure 23). These
results demonstrate that the 1-methylphenanthrene/phenanthrene ratio can
be used to determine whether motor vehicles contributed to the atmospheric
PAH-levels measured with SPMDs. However, when using this indicator it is
important to consider the residential heating systems used in the sampling
area, since the use of fossil fuels for home heating can also cause the
concentrations of methylated PAHs to increase relative to levels of
unsubstituted PAHs. However, the PAH emissions attributable to this source
are negligible in Thailand, and the 1-methylphenanthrene/phenanthrene ratio
can be used as an indicator of traffic’s contribution to the PAHs found in the
SPMDs in such cases.
66
7. Fields of application and sampling strategies
-1
-1
PAH amounts (ng SPMD d )
160
120
80
40
0
0,00
0,25
0,50
0,75
1,00
1-methylphenanthrene/phenanthrene ratio
Figure 23. 1-methylphenanthrene/phenanthrene ratio versus the sum of 14 EPA
PAHs (excluding naphthalene, acenaphthylene) together with benzo[e]pyrene in
SPMDs at sampling sites A-F (Paper I).
In the work reported in Paper IV, the atmospheric levels of the gas phase
nitro-PAHs were measured. Feilberg et al. (1999) analyzed diesel exhaust for
1-NN and 2-NN, and found only 1-NN in the samples, whereas Atkinson et
al. (1987) found that gas phase reactions between naphthalene initiated by
OH radicals produced 1-NN and 2-NN in nearly equal abundance. Thus, the
1-NN to 2-NN concentration ratio can be used to assess whether direct
emissions or gas phase reactions produced the atmospheric concentrations of
nitro-PAHs. The 1-NN and 2-NN ratios found in the SPMD samples
reported in Paper IV were generally close to unity, or between one and two,
which indicate that these nitro-PAHs mainly originated from atmospheric
formation in the gas phase. Thus, this approach can be used to determine
whether direct emissions or gas phase reactions contribute to the levels of 1NN and 2-NN measured with SPMDs.
7.3.1.1 Considerations involved with the use of the individual compound ratios
When comparing the ratios of individual compounds between different
samples, the results can be affected by the purity of the samples, even if
internal standards and selective analytical methods have been used. The
samples in Paper I were re-run on GPC due to interferences with the pyrene
and fluoranthene analysis. The 1-methylphenanthrene/phenanthrene ratios
after the 1st and 2nd GPC analyses displayed a linear relationship (R2 =
0.996), but they were twice as high (slope = 2.14) after the 2nd GPC analyses,
67
7. Fields of application and sampling strategies
demonstrating that the ratios between individual compounds are not
necessarily constant. Photolysis can also affect the detected ratios. The ratios
between 1-NN and 2-NN found in the samples stored for a considerably
longer time before analysis were slightly lower than in the other samples
(Paper IV). Feilberg et al. (1999) have reported that photodegradation
proceeds more rapidly for 1-NN than for 2-NN. Thus, the lower ratios found
in some of the samples could have been caused by higher degradation of 1NN during storage. Unfortunately only 1-nitropyrene and 6-nitrochrysene
were used as internal standards, but the results demonstrate the importance of
using corresponding labelled standards for each of the photosensitive
compounds like nitro-PAHs, to assure the quality of the data. Thus, the ratios
of individual compounds between samples are not always stable for the
reasons mentioned above.
7.3.2 The use of total PAH-patterns
The total PAH-patterns found in the SPMD samples are probably more
stable between samples than the ratios of individual PAH compounds. In
work presented in Paper IV, sampling locations influenced by different
potential sources of PAH pollution were included in order to discern, if
possible, the PAH-pattern of specific sources. As previously described, five
European countries were included in this study. In each country, eight
locations were selected as sampling sites. Participants in each country were
recommended to choose sampling locations (i) in the vicinity and upwind of
at least three different potential sources of PAHs, (ii) at sites in rural areas and
(iii) at one remote site. The pollution situation and the potential sources at
each sampling site are described in detail in Table 1, Paper IV. However, the
degree of air pollution varied between countries and the definitions of
remote, rural and urban areas for each country were somewhat subjective.
Thus, when comparing the data between countries, measured levels were not
always comparable with the subjectively defined pollutant situations of the
respective sampling locations. However, within each country the highest
levels were detected at sites defined as urban, while the lowest levels occurred
at remote and rural sites. According to the site descriptions in Table 1, Paper
IV, the sources likely to be responsible for the high levels of PAHs and nitroPAHs detected at the urban and industrial sites were traffic, metal-, chemical-,
and petrochemical-industries, tar production, coal-fired systems for residential
heating and other types of local combustion systems, which are considered as
general sources of atmospheric PAH-pollution.
68
7. Fields of application and sampling strategies
To visualize the PAH-pattern of each sampling site, and thus attempt to find
correlations between PAH-patterns and specific sources, a PCA model was
fitted to the dataset normalized to the total amount of PAHs found at each
site to exclude the effect of the concentration differences on the results. The
score plot of the first two PCs showed that there were differences in the
PAH-patterns of each site, while there were no differences in the PAHpatterns between countries or laboratories, demonstrating that both the
sampling and the clean-up protocols of the participating research groups were
comparable (Figure 4, Paper IV). However, no correlation between specific
PAH-patterns and the sources described were detected. For example, sites
described as highly influenced by traffic were equally distributed in the score
plot. There are several possible reasons for these results. Firstly, the SPMD
samplers were located about 1 km from the point source and thus, imissions
rather than the emissions of the potential sources were probably measured.
When imissions are measured it is difficult to avoid sources being
confounded, due to the high number of diffusive PAH sources, and the site
descriptions (Table 1, Paper IV) showed that the PAH-levels found at many
sites could originate from multiple sources. Secondly, it was difficult to
control the environmental conditions and the pollution situations during
sampling, and although the intention was to take measurements upwind of
described sources, additional unknown sources could also have influenced the
PAH-levels of each site. Thirdly, volatilization of PAHs from soil could have
affected the results since the samplers were deployed at 1.5 m height. Thus,
there are many diffusive sources that can affect the atmospheric PAH
concentrations, make specific source-patterns of PAHs very difficult to
discern. In addition, the SPMDs sample mainly the gas phase PAHs which
can lower the correlation between sources and PAH data. For instance,
Müller et al. (1998) found that the correlations between traffic and pollutant
data were better for PAHs with 4-rings or more, whereas the correlations for
the gas phase PAHs were generally low.
7.3.3 Determination of residential wood burning as a source of indoor PAH
exposure
Wood burning is considered to be the most important source of PAH
pollution in the air in Sweden, accounting for about 60 percent of the total
PAH emissions, while the contribution of traffic is about 30 percent
(Boström et al., 2002). Thus, wood burning for residential heating is probably
a major source of PAHs in indoor air of Swedish households. In the work
reported in Paper V, the PAH-levels in the indoor environments of some
69
7. Fields of application and sampling strategies
Swedish households were measured with the aim to estimate if residential
wood burning was a source of gas phase PAHs in indoor air. Briefly, SPMDs
(deployed on spiders but not covered by metal umbrellas due to the low air
flow indoors) were placed for two weeks in 16 single-family houses, in the
same residential area, during the winter period 2003 in the small Swedish
town Hagfors, which is situated in mid-Sweden. Nine of the households (w1w9) used wood burning for residential heating, while seven of the households
(nw1-nw7) used other heating systems (electrical or heating pumps). Cigarette
smoking and cooking are other PAH sources that can also contribute to the
PAH pollution of indoor air (Dermentzoglou et al., 2003; Endo et al., 2000).
Thus, households were recommended to avoid smoking indoor during the
sampling period. In addition, it was documented by the participating
households how long time they fried food. Samples were analyzed for 15 of
the 16 EPA PAHs (excluding naphthalene), retene and the seven methylPAHs 2-methylnaphthalene, 1-methylnaphthalene, 2,3-dimethylnaphthalene,
2,3,5,-trimethylnaphthalene, 3-methylphenanthrene, 4,9-methylphenanthrene
and 1-methylphenanthrene.
The total amounts of the 15 EPA PAHs and the seven methyl-PAHs were in
the ranges 22-400 and 16-320 ng ⋅ SPMD-1 ⋅ d-1, respectively (Figure 24).
Thus, the PAH-levels detected were similar or slightly higher than those
found in the SPMDs (range 4.9-170 ng ⋅ SPMD-1 ⋅ d-1) deployed outdoors at
background to urban sites in Sweden during two sampling campaigns, one in
1999 and the other in 2000 (Paper IV). The five highest PAH-levels were
found in SPMDs deployed in houses that used wood burning for residential
heating. In these houses, the PAH-levels detected were three to six times
higher than the lowest levels detected in the other houses. Thus, the PAHlevels in 11 of the 16 houses were similar and rather low and may represent
the background PAH-levels of the indoor air in the area. Four of the five
households (w2, 3, 7 and 8) with the highest levels of gas phase PAHs were
situated in the same block. It is therefore not clear whether the gas-phase
PAHs in the indoor air of these households originated from high PAHemissions indoors, or if PAH-emissions derived from wood burning outdoors
were transferred to the indoor environment via ventilation. However, the
results showed that the use of wood burning for residential heating can
increase the gas phase concentrations of PAHs in indoor environments even
though there were households which used wood burning for residential
heating that had background indoor PAH-levels. One possible explanation
for the differences in measured levels between houses with wood burning
could be that different types (and/or age) of wood was burned, or the
70
7. Fields of application and sampling strategies
operational procedures were different between households. Nevertheless, the
results demonstrate that the use of SPMDs that are not covered by metal
umbrellas, is a highly sensitive sampling approach, and can be used to
determine spatial differences in indoor air concentrations, even though the air
is quite still in indoor environments.
360
-1
-1
PAH amounts (ng SPMD d )
450
270
180
90
nw7
nw6
nw5
nw4
nw3
nw2
nw1
w9
w8
w7
w6
w5
w4
w3
w2
w1
0
Figure 24. Total amounts of 15 EPA PAHs (excluding naphthalene) found in
SPMDs deployed in single-family houses with (black bars) and without (grey bars)
wood burning as residential heating system (modified from Paper V).
Retene (1-methyl-7-propylphenanthrene) has been suggested as an indicator
of atmospheric PAH pollution from wood burning (Ramdahl, 1983;
Hawthorne et al. 1988). Retene was found in all samples, indicating that it was
present in the indoor environments, but the amounts were low and ranged
from 0.3-1.4 ng ⋅ SPMD-1 ⋅ d-1. High retene levels were detected in houses
which used wood burning as heating system, but the highest concentration of
retene was detected in a house which was not using this type of heating. Thus,
the correlation between retene data and the use of residential wood burning
was unclear. A possible reason for these results could be that retene is more
abundant in emissions from burning of softwood (like coniferious wood) as
compared to hardwood (McDonald et al., 2000). Unfortunately, the type of
wood burned in this study was not documented, and the utility of retene to
detect wood burning as the source of gas phase PAHs in indoor air cannot be
clarified from the data obtained in the work presented in Paper V.
71
7. Fields of application and sampling strategies
7.4 Estimation of the consequences of gas phase PAH exposure for
plants
Many studies have found that POPs may accumulate from the atmosphere
into plants (e.g., Voutsa and Samara, 1998; Ockenden et al., 1998b;
Kipopoulou et al., 1999; Müller et al., 2001; Smith et al., 2001; Kylin and
Sjödin, 2003; Hellström et al., 2004). The plant uptake from the air occurs
mainly through gas deposition or wet/dry particle deposition. Gas phase
PAHs are generally adsorbed to the plant surface and/or accumulated via the
stomata in the cuticle of the plants, whereas particle-associated PAHs are
mainly retained on the plant surface (Smith and Jones, 2000). In the work
described in Paper I, the uptake of PAHs into the water-surface living plant
water spinach (Ipomoea Aquatica), and the gas phase concentration of PAHs,
was sampled in parallel at three different locations in rural (B), semiurban (C)
and urban (D) area in the Bangkok region, Thailand. Water spinach is a
perennial, semiaquatic plant that occurs, both wild and cultivated, in the
southern Asia and other subtropical areas (Figure 25). Its growth rate is fast
and the young shoot, about the upper 30 cm of the plant, is a highly-regarded
food that is frequently consumed by the inhabitants of the Bangkok region,
Thailand. Since water spinach is often cultivated in rivers and canals in
polluted areas, the plant is a potential source of PAH-exposure for humans.
Thus, this plant species was selected, even though it is both air and waterliving, which complicate the data interpretation by adding water as a possible
exposure route for the plant.
72
7. Fields of application and sampling strategies
Figure 25. Illustration of Ipomoea Aquatica provided by IFAS, Centre for aquatic
plants, University of Florida, Gainesville, 1990. The plant has a long, thin stem that
generally grows floating on the water surface with horizontally oriented leaves,
mostly
held
slightly
above
the
water
surface
(obtained
from
http://aquat1.ifas.ufl.edu/ipaqpic.html).
At each of the three locations B, C and D, one SPMD sampler was deployed
at the river side for three weeks. At the time of retrieval of the SPMD
samplers, 20 – 25 cm of the upper part of each plant (the edible part) was cut
off, wrapped in aluminium foil and transported to the Asian Institute of
Technology (AIT), Bangkok, Thailand. These plant shoots were three to ten
days old, and had mostly been growing above the water surface. In the
laboratory, the plant samples were rinsed in tap water followed by distilled
water on the day of collection, to remove particles from the plant surface.
Shoots from ten plants were pooled per sample, and triplicate samples were
obtained at each site. The samples were weighed (wet weight (ww)), wrapped
in aluminium foil and stored at -18 °C. The plant and the SPMD samples
were analysed for 15 of the 16 EPA PAHs (excluding acenaphthylene) plus
benzo[e]pyrene. The PAH amounts found in the SPMDs were converted to
gas phase concentrations using the approach discussed in 4.5.1.
The total average PAH concentrations (n = 3) of the plant samples ranged
from 4.2 to 9.6 µg ⋅ g-1 ww, with two times higher levels at site D than at sites
B and C. In order to compare these results with data obtained in other
studies, the dry weight (dw) of the plant was estimated to be 10 percent of the
73
7. Fields of application and sampling strategies
ww. The recalculated average plant concentrations ranged from 42 to 96 µg ⋅
kg-1 dw and were in the same range as concentrations detected for the 16
EPA PAHs in vegetables growing in an industrial area of northern Greece
(Voutsa and Samara, 1998; Kipopoulou et al., 1999).
4
DahA
IcdP
IcdP
BghiP
BaP
BaP
DahA
Py
BbF, BkF
Py
BbF, BkF
Chr
Fluo
Ant
Phe
Fl
Ace
0
Chr
2
Fluo
a)
6
BaA
Plant concentration (ug kg-1 ww)
8
-3
Air concentration (ng m )
12
6
BghiP
Ant
Phe
Fl
0
BaA
3
Ace
b)
9
Figure 26. Individual PAH concentrations in a) the plants and b) the gas phase at
sampling sites B (●), C (▲) and D (■) (Paper I).
The major individual PAHs found in the gas phase and the plants were
fluorene, phenanthrene, fluoranthene and pyrene (Figure 26). However,
PAHs with three rings (MW < 202) were predominant in the gas phase, while
the PAH-profiles found in the plants were slightly different with levels of 4ring PAHs (MW < 202) that were higher or similar to the levels of 3-ring
PAHs. Thus, gas phase and particle-bound PAHs were accumulated in plants
to a similar degree, indicating that particle phase exposure contributed to the
levels found in the plants. Another explanation for the differences in the
PAH-profiles between the gas phase sampled with SPMDs, and the plants is
74
7. Fields of application and sampling strategies
that the plants could have taken up additional amounts of the PAHs from the
water. However, possible PAH exposure from the water to the plants was not
investigated in this study. Furthermore, the concentration of pyrene in the
plants (higher levels at site B than C) was not correlated with the traffic
intensity at the sampling site. Thus, sources other than traffic as well as other
exposure routes than the gas phase like particle and water exposure could
have contributed to the higher concentration of pyrene (a 4-ring PAH) in the
plants compared to the SPMDs. However, the similarities in the SPMD and
plant profiles of 3-ring PAHs demonstrate that the gas phase exposure
contributed to the amounts of PAHs accumulated in the plants.
75
76
8. Conclusions and future research possibilities
8. CONCLUSIONS AND FUTURE RESEARCH POSSIBILITIES
The work described in this thesis tested the use of SPMDs as integrative tools
for monitoring gas phase concentrations of the nonpolar aromatic pollutants
PCBs, PAHs, alkyl-PAHs and nitro-PAHs. Conclusions drawn from this
work are listed below, and further research possibilities are suggested.
The uptake and release in SPMDs of PAHs and PCBs with log KOA values >
7.9 increase at high wind-speeds (Papers II and III). These results
demonstrate that the uptake of most nonpolar aromatic compounds is
controlled by the boundary layer at the membrane-air interface. However, the
wind-speeds/turbulence in the field will be more moderate and less variable
between locations, and thus, the wind effect will be generally low. In future
research, the effect of temperature, UV light and particle deposition should
be examined. The processes involved in the effect of cold temperature are
especially important to understand.
The use of a metal umbrella to shelter the SPMDs decreases the uptake and
release of PAHs and PCBs at high wind-speeds and/or turbulence, and thus
reduces the wind effect (Paper III). In addition, the uptake of particleassociated PAHs increases in unprotected SPMDs, indicating that the metal
umbrellas reduce the effect of particle deposition (Paper III). However, the
sampler design used in the work underlying this thesis is probably not optimal
for reducing the effect of reflected UV light. In future research the design of
the metal umbrella should be improved by testing, for instance, different dark
materials to coat the inside of the metal umbrella.
The PRC approach has a high ability to assess the effect of high wind (Papers
II and III). However, analytical interferences might reduce the precision of
the PRCs (Papers II and III), and it is critical that robust analytical quality
control is applied. The effects of particles will probably not be assessed by the
use of PRCs (Paper III). The use of the PRC-approach to predict the effect of
temperature and photolysis has not been tested. The air temperatures can be
highly variable and its effect is quite complex and rather unknown. Thus, the
use of the PRC-approach to predict the temperature effect is a very important
task to study in future research. The use of photosensitive compounds to
predict the effect of photolysis is another future research issue.
77
8. Conclusions and future research possibilities
SPMDs are efficient samplers of the gas phase concentrations of nonpolar
aromatic compounds, and are able to determine the spatial distribution at
local scales (Paper I and IV). If uncovered SPMDs are used, they can be
highly sensitive samplers, and can be used to determine spatial differences
under quite still air conditions, such as those in indoor environments (Paper
V). In addition, the use of SPMDs, which do not require electricity, make
sampling possible at remote and rural areas where the infrastructure is limited,
like the countryside in Thailand (Paper I). SPMDs are also suitable for
integrative sampling of the atmospheric distribution of nonpolar aromatic
compounds on a continental scale (Paper IV).
The sources of pollution can also be determined with SPMDs (Papers I, IV
and V). The 1-methylphenanthrene/phenanthrene ratio is a potential
indicator of the contribution of traffic to the gas phase concentrations
measured with SPMDs (Paper I). Another approach to determine the
pollution source of PAHs is to discern the total PAH-patterns of specific
sources. However, SPMDs cannot be used to discern the total PAH-pattern
of specific sources, probably because of the high number of diffusive PAH
sources influencing the imission measurements (Paper IV). In addition,
SPMDs can be used to detect wood burning for residential heating as a
possible source of indoor PAH exposure (Paper V), and to determine the
importance of the gas phase exposure route to the uptake of PAHs in plants
(Paper I).
SPMDs have several advantages in comparison with conventionally used
HiVols, including their integrative capacity over long times, reduced costs,
and no need of specialist operators, maintenance or power supply during
sampling. Thus, the use of SPMDs can increase the monitoring frequency, the
geographical distribution of the measurements, and the number of sampling
sites, used in air monitoring programs of nonpolar aromatic compounds.
However, calibration data of SPMDs are limited, and spatial differences are
often only semi-quantitatively determined by comparing amounts and profiles
in the SPMDs, which have limited their use in air monitoring programs. In
future work, it is important to ensure that SPMDs are properly sheltered,
PRCs are used in the sampling protocols, and that calibrated RS data, or the
SPMD-air partition data, of specific compounds are further developed to
make determination of TWA concentrations possible.
78
9. Acknowledgements
9. ACKNOWLEDGEMENTS
Så har jag kommit fram till den del av avhandlingen då jag vill tacka alla er som hjälpt mig
under min tid i Umeå som doktorand och peppat mig att simma ända in i kaklet!!
I am grateful for financial support of this work by the European Commission INCO
program and the Austrian ministry of education, research and culture. Tack också Helge
Ax:son Johnsons stiftelse, Knut och Alice Wallenberg-stiftelsen, the Swedish International
Development Cooperation Agency (SIDA), samt Stiftelsen JC Kempes minnes-stipendiefond
för ekonomiskt stöd av projekt, resor och konferenser.
Tack alla härliga människor på miljökemi som skapar en bra och avslappnad stämmning
på jobbet. Speciellt nöjd är jag med superfreda´ som innehåller både innebandy (även om det
var ett tag sen jag var med) och gofika, tack för det allihopa! Sen finns det ju alltid några som
man vill tacka lite extra. Tack min handledare Pekka för ditt lugn, stöd, dina roliga
projektideér, och att du alltid tar dig tid att diskutera forskning med mig. Tack också för alla
resor du har tagit med mig på, och för att du har presenterat mig för många duktiga och
inspirerande forskare världen över. Bilresan genom klippigabergen, Utah, är ett av många fina
minnen från våra jobbresor! Tack också Mats för din handledning, att du stöttat mig i tuffa
tider och din konstruktiva kritik av mitt arbete och min avhandling. Tack Patrik för att du
lärde mig massor under mitt ex-arbete, och Lena, för att du alltid bryr dig om andra och
hjälper och stöttar när det behövs. Tack Fia för att du är så go´ och glad, och att du får mig
att skratta! Tack Eva för att du drog med mig till Thailand, och för alla pratstunder om
forsket och livet, ditt stöd har varit guld värt! Tack Lijana för din stora omtanke, glada humör
och för att du är super att resa med. ”Din hjälp” innan min muntliga presentation i Salt Lake
City gjorde susen ;- )! Thanks also to James Huckins, Columbia, USA, and Bo Strandberg,
Gothenburg, for sharing your superb SPMD knowledge with me. Keep up the good work!
Tack också Leif Lindgren, Älvsbyn, för dina fina illustrationer.
Tack alla ni som nu har flyttat men förgyllde min studietid i Umeå med roliga fester,
middagar och annat skoj, tack för den härliga tiden! Tack alla vänner här i Umeå för att ni
finns och gör var dag kul och gjort doktorerandet lättare. Vad vore en dag utan en TV-kväll
med sport, en torsdag utan 24, en kväll utan en runda på IKSU eller en helg utan en skogstur
med cykeln?! Tack Älvsbybrudarna för att vi hållit ihop i alla år, allt kul vi har haft
tillsammans och för att ni alltid ger mig nya perspektiv på livet, vilket behövs som doktorand!
Tack tjejerna i mailgruppen, vad vore en måndag utan en brevlåda full med härliga historia
och skvaller. Hoppas jag inte har tråkat ut er på slutet med ändlösa klago-mail. Tjejhelgen i
höstmörkret gjorde iallafall super, ser fram emot nästa!
För att använda ett slitet uttryck vill jag sist men inte minst tacka min underbara familj.
Men hur ska jag kunna skriva vad ni har betytt för mig på bara några rader?! Det går inte helt
enkelt... Mamma och Pappa, ni har alltid stöttat och uppmuntrat mig något helt enormt. Utan
er hade jag inte varit den jag är, eller nått dit jag är idag! Som ni har peppat mig, ända fram till
tryckdag! Tack också mormor för att du alltid brydda dig om ”små-flickorna i Ume”, du finns
alltid i mina tankar. Och sen Syrris, hur mycket kul och fint har inte vi haft tillsammans?!
Tack för allt, du är världens bästa syster! Och så Anton, min älskade sambo, med dig vid min
sida är ingenting omöjligt! Du har peppat, hjälpt och stöttat mig något helt fantastiskt. Tack
för att du är så fin!
79
Det går inte bromsa sig ur en uppförsbacke.
Illustration by Leif Lindgren, Älvsbyn
80
References
REFERENCES
Aas, W., and Breivik, K., 2004. Heavy metals and POP measurements, 2002.
EMEP/CCC-Report 7/2004. Norwegian institute for air research, Kjeller, Norway.
(see also http://www.nilu.no/projects/ccc/reports.html)
Aga, E., Samoli, E., Touloumi, G., Anderson, H.R., Cadum, E., Forsberg, B.,
Goodman, P., Goren, A., Kotesovec, F., Kriz, B., Macarol-Hiti, M., Medina, S.,
Paldy, A., Schindler, C., Sunyer, J., Tittanen, P., Wojtyniak, B., Zmirou, D., Schwartz,
J., and Katsouyanni, K., 2003. Short-term effects of ambient particles on mortality in
the elderly: results from 28 cities in the APHEA2 project. European Respiratory
Journal. 21, 28S-33S.
Anonymous, 2001. EMEP manual for sampling and chemical analysis. Report
1/1995, Revision 1/2001. Norwegian Institute for Air Research, Kjeller, Norway.
(see also http://www.nilu.no/projects/ccc/manual/index.html updated February,
2004 by Wenche Aas ([email protected]))
Anonymous, 2002a. Ambient air quality-diffusive samplers for the determination of
concentrations of gases and vapours - Requirements and test methods- Part 1:
General requirements. EN 13528-1:2002. European Committee for standardization,
Brussels, Belgium.
Anonymous, 2002b. Ambient air quality-diffusive samplers for the determination of
concentrations of gases and vapours - Requirements and test methods- Part 2:
Specific requirements and test methods. EN 13528-2:2002. European Committee for
standardization, Brussels, Belgium.
Anonymous, 2003. Ambient air quality-diffusive samplers for the determination of
concentrations of gases and vapours - Requirements and test methods- Part 3: Guide
to selection, use and maintenance. EN 13528-3:2003. European Committee for
standardization, Brussels, Belgium.
Arey, J., Zielinska, B., Winer, A.M., Ramdahl, T., and Pitts Jr, J.N., 1986. The
formation of nitro-PAH from gas-phase reactions of fluoranthene and pyrene with
the OH radical in the present of NOx. Atmospheric Environment. 20, 2339-2345.
Atkinson, R., and Arey, J., 1994. Atmospheric chemistry of gas-phase polycyclic
aromatic hydrocarbons - Formation of atmospheric mutagens. Environmental
Health Perspectives. 102, 117-126.
81
References
Atkinson, R., Arey, J., Zielinska, B., Pitts Jr, J.N., and Winer, A.M., 1987. Evidence
for the transformation of polycyclic organic matter in the atmosphere. Atmospheric
Environment. 21, 2261-2264.
Ballschmitter, K., and Zeller, M., 1980. Analysis of polychlorinated biphenyls (PCBs)
by glass capillary gas chromatography. Composition of technical Aroclor- and
Clophen-PCB mixtures. Fresenius Journal of Analytical Chemistry. 302, 20-31.
Balmer, M.E., Poiger, T., Droz, C., Romanin, K., Bergqvist, P-A., Müller, M.D., and
Buser, H.R., 2004. Occurrence of methyl triclosan, a transformation product of the
bactericide triclosan, in fish from various lakes in Switzerland. Environmental
Science & Technology. 38, 390-395.
Bamford, H.A., Bezabeh, D.Z., Schantz, M.M., Wise, S.A., and Baker, J.E., 2003.
Determination and comparison of nitrated-polycyclic aromatic hydrocarbons
measured in air and diesel particulate reference materials. Chemosphere. 50, 575-587.
Bartkow, M.E., Hawker, D.W., Kennedy, K.E., and Müller, J.F., 2004. Characterizing
uptake kinetics of PAHs from the air using polyethylene-based passive air samplers
of multiple surface area-to-volume ratios. Environmental Science & Technology. 38,
2701-2706.
Bartkow, M.E., Huckins, J.N., and Müller, J.F., 2004. Field-based evaluation of
semipermeable membrane devices (SPMDs) as passive air samplers of polyaromatic
hydrocarbons (PAHs). Atmospheric Environment. 38, 5983-5990.
Berglund, M., Vahter, M., and Bylin, G., 1992. Measurement of Personal Exposure
to NO2 in Sweden - Evaluation of A Passive Sampler. Journal of Exposure Analysis
and Environmental Epidemiology. 2, 295-307.
Bergqvist, P-A., Strandberg, B., Ekelund, R., Rappe, C., and Granmo, Å, 1998.
Temporal monitoring of organochlorine compounds in seawater by semipermeable
membranes following a flooding episode in western Europe. Environmental Science
& Technology. 32, 3887-3892.
Booij, K., Hofmans, H.E., Fischer, C.V., and van Weerlee, E.M., 2003. Temperaturedependent uptake rates of nonpolar organic compounds by semipermeable
membrane devices and low-density polyethylene membranes. Environmental Science
& Technology. 37, 361-366.
Booij, K., Sleiderink, H.M., and Smedes, F., 1998. Calibrating the uptake kinetics of
semipermeable membrane devices using exposure standards. Environmental
Toxicology and Chemistry. 17, 1236-1245.
82
References
Booij, K., and van Drooge, B.L., 2001. Polychlorinated biphenyls and
hexachlorobenzene in atmosphere, sea-surface microlayer, and water measured with
semi-permeable membrane devices (SPMDs). Chemosphere. 44, 91-98.
Boström, C.E., Gerde, P., Hanberg, A., Jernström, B., Johansson, C., Kyrklund, T.,
Rannug, A., Tornqvist, M., Victorin, K., and Westerholm, R., 2002. Cancer risk
assessment, indicators, and guidelines for polycyclic aromatic hydrocarbons in the
ambient air. Environmental Health Perspectives. 110, 451-488.
Bosveld, A.T., Nieboer, R., de Bont, A., Mennen, J., Murk, A.J., Feyk, L.A., Giesy,
J.P., and van den Berg, M., 2000. Biochemical and developmental effects of dietary
exposure to polychlorinated biphenyls 126 and 153 in common tern chicks (Sterna
hirundo). Environmental Toxicology and Chemistry. 19, 719-730.
Bytnerowicz, A., Godzik, B., Fraczek, W., Grodzinska, K., Krywult, M., Badea, O.,
Barancok, P., Blum, O., Cerny, M., Godzik, S., Mankovska, B., Manning, W.,
Moravcik, P., Musselman, R., Oszlanyi, J., Postelnicu, D., Szdzuj, J., Varsavova, M.,
and Zota, M., 2002. Distribution of ozone and other air pollutants in forests of the
Carpathian Mountains in central Europe. Environmental Pollution. 116, 3-25.
Cao, X.L., and Hewitt, C.N., 1993. Evaluation of Tenax-Gr Adsorbent for the
passive sampling of volatile organic compounds at low concentrations. Atmospheric
Environment Part A-General Topics. 27, 1865-1872.
Chen, J., Xue, X., Schramm, K-W., Quan, X., Yang, F., and Kettrup, A., 2002.
Quantitative structure-property relationships for octanol-air partition coefficients of
polychlorinated biphenyls. Chemosphere. 48, 535-544.
Chen, B.H., Hong, C.J., and Kan, H.D., 2004. Exposures and health outcomes from
outdoor air pollutants in China. Toxicology. 198, 291-300.
Chiou C.T., 1985. Partition coefficients of organic compounds in lipid-water systems
and correlations with fish bioconcentration factor. Environmental Science &
Technology. 19, 57-62.
Dermentzoglou, M., Manoli, E., Voutsa, D., and Samara, C., 2003. Sources and
patterns of polycyclic aromatic hydrocarbons and heavy metals in fine indoor
particulate matter of Greek houses. Fresenius Environmental Bulletin. 12, 15111519.
Dimashki, M., Harrad, S., and Harrison, R.M., 2000. Measurements of nitro-PAH in
the atmospheres of two cities. Atmospheric Environment. 34, 2459-2469.
83
References
Durant, J.L., Busby, W.F., Lafleur, A.L., Penman, B.W., and Crespi, C.L., 1996.
Human cell mutagenicity of oxygenated, nitrated and unsubstituted polycyclic
aromatic hydrocarbons associated with urban aerosols. Mutation Research-Genetic
Toxicology. 371, 123-157.
Dusek, B., Hajskova, J., and Kocourek, V., 2002. Determination of nitrated
polycyclic aromatic hydrocarbons and their precursors in biotic matrices. Journal of
Chromatography A. 982, 127-143.
Echols, K.R., Gale, R.W., Schwartz, T.R., Huckins, J.N., Williams, L.L., Meadows,
J.C., Morse, D., Petty, J.D., Orazio, C.E., and Tillitt, D.E., 2000. Comparing
polychlorinated biphenyl concentrations and patterns in the Saginaw River using
sediment, caged fish, and semipermeable membrane devices. Environmental Science
& Technology. 34, 4095-4102.
Ellis, G.S., Huckins, J.N., Rostad, C.E., Schmitt, C.J., Petty, J.D., and Maccarthy, P.,
1995. Evaluation of lipid-containing semipermeable-membrane devices for
monitoring organochlorine contaminants in the Upper Mississippi River.
Environmental Toxicology and Chemistry. 14, 1875-1884.
Endo, O., Koyano, M., Mineki, S., Goto, S., Tanabe, K., Yajima, H., Ishii, T., and
Matsushita, H., 2000. Estimation of indoor air PAH concentration increases by
cigarette, incense-stick, and mosquito-repellent-incense smoke. Polycyclic Aromatic
Compounds. 21, 261-272.
Eriksson, L., Johansson, E., Kettaneh-Wold, N., and Wold, S., 1999. Introduction to
multi- and megavariate data analysis using projection methods (PCA and PLS).
Umetrics AB, Umeå, Sweden.
Feilberg, A., Kamens, R.M., Strommen, M.R., and Nielsen, T., 1999. Modeling the
formation, decay, and partitioning of semivolatile nitro-polycyclic aromatic
hydrocarbons (nitronaphthalenes) in the atmosphere. Atmospheric Environment. 33,
1231-1243.
Forsberg, B., Stjernberg, N., Linne, R., Segerstedt, B., and Wall, S., 1998. Daily air
pollution levels and acute asthma in southern Sweden. European Respiratory
Journal. 12, 900-905.
Gale, R.W., 1998. Three-compartment model for contaminant accumulation by
semipermeable membrane devices. Environmental Science & Technology. 32, 22922300.
84
References
Giesy, J.P. and Kannan, K., 1998. Dioxin-like and non-dioxin-like toxic effects of
polychlorinated biphenyls (PCBs): Implications for risk assessment. Critical Reviews
in Toxicology. 28, 511-569.
Gorecki, T. and Namiesnik, J., 2002. Passive sampling. Trac-Trends in Analytical
Chemistry. 21, 276-291.
Granmo, Å., Ekelund, R., Berggren, M., Brorström-Lunden, E., and Bergqvist, P-A.,
2000. Temporal trend of organochlorine marine pollution indicated by
concentrations in mussels, semipermeable membrane devices, and sediment.
Environmental Science & Technology. 34, 3323-3329.
Harner, T., and Bidleman, T.F., 1996. Measurements of octanol-air partition
coefficients for polychlorinated biphenyls. Journal of Chemical and Engineering
Data. . 41, 895-899.
Harner, T., and Bidleman, T.F., 1998. Measurement of octanol-air partition
coefficients for polycyclic aromatic hydrocarbons and polychlorinated naphthalenes.
Journal of Chemical and Engineering Data. 43, 40-46.
Harner, T., Farrar, N.J., Shoeib, M., Jones, K.C., and Gobas, F.A., 2003.
Characterization of polymer-coated glass as a passive air sampler for persistent
organic pollutants. Environmental Science & Technology. 37, 2486-2493.
Hawthorne, S.B., Miller, D.J, Barkley, R.M., and Krieger, M.S., 1988. Identification
of methoxylated phenols as candidate tracers for atmospheric wood smoke
pollution. Environmental Science & Technology. 22, 1191-1196.
Hellström, A., Kylin, H., Strachan, W.M., and Jensen, S., 2004. Distribution of some
organochlorine compounds in pine needles from Central and Northern Europe.
Environmental Pollution. 128, 29-48.
Herve, S., Paukku, R., Paasivirta, J., Heinonen, P., and Södergren, A., 1991. Uptake
of organochlorines from lake Water by hexane-filled dialysis membranes and by
mussels. Chemosphere. 22, 997-1001.
Huckins, J., Tubergen, M.W., and Manuweera, G.K., 1990. Semipermeable
membrane devices containing model lipid: a new approach to monitoring the
bioavailability of lipophilic contaminants and estimating their bioconcentration
potential. Chemosphere. 20, 533-552.
Huckins, J. N., 2004. Personal communication. United States Geological Survey’s,
Columbia Environmental Research centre, Columbia, USA.
85
References
Huckins, J.N., Lebo, J.A., Tubergen, M.W., Manuweera, G.K., Gibson, V.L., and
Petty, J.D., 1992. Binary concentration and recovery process. U.S. Patent, 5,098,573.
Huckins, J.N., Manuweera, G.K., Petty, J.D., Mackay, D., and Lebo, J.A., 1993.
Lipid-containing semipermeable-membrane devices for monitoring organic
contaminanst in water. Environmental Science & Technology. 27, 2489-2496.
Huckins, J.N., Petty, J.D., Lebo, J.A., Almeida, F.V., Booij, K., Alvarez, D.A., Clark,
R.C., and Mogensen, B.B., 2002a. Development of the permeability/performance
reference compound approach for in situ calibration of semipermeable membrane
devices. Environmental Science & Technology. 36, 85-91.
Huckins, J.N., Petty, J.D., Orazio, C.E., Lebo, J.A., Clark, R.C., Gibson, V.L., Gala,
W.R., and Echols, K.R., 1999. Determination of uptake kinetics (sampling rates) by
lipid-containing semipermeable membrane devices (SPMDs) for polycyclic aromatic
hydrocarbons (PAHs) in water. Environmental Science & Technology. 33, 39183923.
Huckins, J.N., Petty, J.D., Prest, H.F., Clark, R.C., Alvarez, D.A., Orazio, C.E., Lebo,
J.A., Cranor, W.L., and Johnson, B.T., 2002b. A guide for the use of semipermeable
membrane devices (SPMDs) as samplers of waterborne hydrophobic organic
contaminants. American Petroleum Institute, Washington, USA.
Huckins, J.N., Petty, J.D., Zajicek, J.L., and Gibson, V.L., 1995. Device for the
removal and concentration of organic compounds from the atmosphere. U.S. Patent,
5,395,426.
Ikonomou, M.G., Rayne, S., Fischer, M., Fernandez, M.P., and Cretney, W., 2002.
Occurrence and congener profiles of polybrominated diphenyl ethers (PBDEs) in
environmental samples from coastal British Columbia, Canada. Chemosphere. 46,
649-663.
Isidori, M., Ferrara, M., Lavorgna, M., Nardelli, A., and Parrella, A., 2003. In situ
monitoring of urban air in Southern Italy with the tradescantia micronucleus
bioassay and semipermeable membrane devices (SPMDs). Chemosphere. 52, 121126.
Jackson, J.E., 1991. A users guide to principal components. John Wiley & Sons, New
York, USA.
Jaward, F.M., Farrar, N.J., Harner, T., Sweetman, A.J., and Jones, K.C., 2004a.
Passive air sampling of PCBs, PBDEs, and organochlorine pesticides across Europe.
Environmental Science & Technology. 38, 34-41.
86
References
Jaward, F.M., Farrar, N.J., Harner, T., Sweetman, A.J., and Jones, K.C., 2004b.
Passive air sampling of polycyclic aromatic hydrocarbons and polychlorinated
naphthalenes across Europe. Environmental Toxicology and Chemistry. 23, 13551364.
Johansson L., 2004. Personal communication. Applied physics and electronics, Umeå
University, Umeå, Sweden. (See also www.tfe.umu.se/weathr/arkiv.asp, 2004-10-25.)
Johnson, G.D., 1991. Hexane-filled dialysis bags for monitoring organic
contaminants in water. Environmental Science & Technology. 25, 1897-1903.
Kannan, K., Blankenship, A.L., Jones, P.D., and Giesy, J.P., 2000. Toxicity reference
values for the toxic effects of polychlorinated biphenyls to aquatic mammals. Human
and Ecological Risk Assessment. 6, 181-201.
Kipopoulou, A.M., Manoli, E., and Samara, C., 1999. Bioconcentration of polycyclic
aromatic hydrocarbons in vegetables grown in an industrial area. Environmental
Pollution. 106, 369-380.
Krochmal, D., and Kalina, A., 1997. A method of nitrogen dioxide and sulphur
dioxide determination in ambient air by use of passive samplers and ion
chromatography. Atmospheric Environment. 31, 3473-3479.
Kruså, M., Bellander, T., and Nilsson, M., 2004. Cancerframkallande ämnen i
tätortsluft Stockholm 2002/2003. Report for national environmental monitoring.
Swedish Environmental Protection Agency, Stockholm, Sweden.
Kylin, H., and Sjödin, A., 2003. Accumulation of airborne hexachlorocyclohexanes
and DDT in pine needles. Environmental Science & Technology. 37, 2350-2355.
Lebo, J.A., Zajicek, J.L., Huckins, J.N., Petty, J.D., and Peterman, P. H., 1992. Use of
semipermeable-membrane devices for in situ monitoring of polycyclic aromatichydrocarbons in aquatic environments. Chemosphere, 25, 697-718.
Lehndorff, E., and Schwark, L., 2004. Biomonitoring of air quality in the Cologne
Conurbation using pine needles as a passive sampler - Part II: polycyclic aromatic
hydrocarbons (PAH). Atmospheric Environment. 38, 3793-3808.
Letcher, R.J., Lemmen, J.G., van der Burg, B., Brouwer, A., Bergman, A., Giesy, J.P.,
and van den Berg, M., 2002. In vitro antiestrogenic effects of aryl methyl sulfone
metabolites of polychlorinated biphenyls and 2,2-bis(4-chlorophenyl)-1,1dichloroethene on 17 beta-estradiol-induced gene expression in several bioassay
systems. Toxicological Sciences. 69, 362-372.
87
References
Lim, L.H., Harrison, R.M., and Harrad, S., 1999. The contribution of traffic to
atmospheric concentrations of polycyclic aromatic hydrocarbons. Environmental
Science & Technology. 33, 3538-3542.
Lindström, A., Buerge, I.J., Poiger, T., Bergqvist, P-A, Müller, M.D., and Buser,
H.R., 2002. Occurrence and environmental behavior of the bactericide triclosan and
its methyl derivative in surface waters and in wastewater. Environmental Science &
Technology. 36, 2322-2329.
Lohmann, R., Corrigan, B.P., Howsam, M., Jones, K.C., and Ockenden, W.A., 2001.
Further developments in the use of semipermeable membrane devices (SPMDs) as
passive air samplers for persistent organic pollutants: Field application in a spatial
survey of PCDD/Fs and PAHs. Environmental Science & Technology. 35, 25762582.
Lohmann, R., Harner, T., Thomas, G.O., and Jones, K.C., 2000. A comparative
study of the gas-particle partitioning of PCDD/Fs, PCBs, and PAHs. Environmental
Science & Technology. 34, 4943-4951.
Luellen, D.R., and Shea, D., 2002. Calibration and field verification of
semipermeable membrane devices for measuring polycyclic aromatic hydrocarbons
in water. Environmental Science & Technology. 36, 1791-1797.
Mackay D, Shiu.W.Y., and Ma K.C., 1992. Illustrated Handbook of physicalchemical properties and environmental fate for organic chemicals (Volume II). Lewis
Publishers, Chelsea, Michigan, USA.
McDonald, R.D., Zielinska, B., Fujita, E.M., Sagebiel, J.C., Chow, J.C., and Watson,
J.G., 2000. Fine particle and gaseous emission rates from residential wood
combustion. Environmental Science & Technology. 34, 2080-2091.
Meadows, J.C., Echols, K.R., Huckins, J.N., Borsuk, F.A., Carline, R.F., and Tillitt,
D.E., 1998. Estimation of uptake rate constants for PCB congeners accumulated by
semipermeable membrane devices and brown trout (Salmo trutta). Environmental
Science & Technology. 32, 1847-1852.
Meijer, S.N., Ockenden, W.A., Steinnes, E., Corrigan, B.P., and Jones, K.C., 2003.
Spatial and temporal trends of POPs in Norwegian and UK background air:
Implications for global cycling. Environmental Science & Technology. 37, 454-461.
Modig, L., Forsberg, B., Hagenbjörk-Gustafsson, A., Järvholm, B., Levin, J-O,
Lindahl, R., Rhe´n, M., Segerstedt, B., Sundgren, M., Sunesson, A-L, and BrorströmLunden, E., 2002. Cancerframkallande ämnen tätortsluft - Exponering och halter i
88
References
Umeå 2001. Report for national environmental monitoring. Swedish Environmental
Protection Agency, Stockholm, Sweden.
Müller, J.F., Hawker, D.W., and Connell, D.W., 1998. Polycyclic aromatic
hydrocarbons in the atmospheric environment of Brisbane, Australia. Chemosphere.
37, 1369-1383.
Müller, J.F., Hawker, D.W., Connell, D.W., Komp, P., and McLachlan, M.S., 2000.
Passive sampling of atmospheric SOCs using tristearin-coated fibreglass sheets.
Atmospheric Environment. 34, 3525-3534.
Müller, J.F., Hawker, D.W., McLachlan, M.S., and Connell, D.W., 2001. PAHs,
PCDD/Fs, PCBs and HCB in leaves from Brisbane, Australia. Chemosphere. 43,
507-515.
Nielsen, T., 1984. Reactivity of polycyclic aromatic hydrocarbons towards nitrating
species. Environmental Science & Technology. 18, 157-163.
Nielsen, T., 1996. Traffic contribution of polycyclic aromatic hydrocarbons in the
center of a large city. Atmospheric Environment. 30, 3481-3490.
Ockenden, W.A., Corrigan, B.P., Howsam, M., and Jones, K.C., 2001. Further
developments in the use of semipermeable membrane devices as passive air
samplers: Application to PCBs. Environmental Science & Technology. 35, 45364543.
Ockenden, W.A., Prest, H.F., Thomas, G.O., Sweetman, A., and Jones, K.C., 1998a.
Passive air sampling of PCBs: Field calculation of atmospheric sampling rates by
triolein-containing semipermeable membrane devices. Environmental Science &
Technology. 32, 1538-1543.
Ockenden, W.A., Steinnes, E., Parker, C., and Jones, K.C., 1998b. Observations on
persistent organic pollutants in plants: Implications for their use as passive air
samplers and for POP cycling. Environmental Science & Technolog. 32, 2721-2726.
Ockenden, W.A., Sweetman, A.J., Prest, H.F., Steinnes, E., and Jones, K.C., 1998c.
Toward an understanding of the global atmospheric distribution of persistent
organic pollutants: The use of semipermeable membrane devices as time-integrated
passive samplers. Environmental Science & Technology. 32, 2795-2803.
Opperhuizen A., Velde E.W., Gobas F.A., Liem, A.K., and Steen J.M., 1985.
Relationship between bioconcentration in fish and steric factors of hydrofobic
chemicals. Chemosphere. 14, 1871-1896.
89
References
Orazio, C.E., Lebo, J.A., Petermann, P.H., Meadows, J.C., Huckins, J.N., and Petty,
J., 2002. Potential for photodegradation of contaminants during SPMD sampling.
Poster (p196). SETAC 23rd annual meeting in north America, November 16-20, Salt
Lake City, Utah, USA.
Petty, J.D., Huckins, J.N., Alvarez, D.A., Brumbaugh, W.G., Cranor, W.L., Gale,
R.W., Rastall, A.C., Jones-Lepp, T.L., Leiker, T.J., Rostad, C.E., and Furlong, E.T.,
2004. A holistic passive integrative sampling approach for assessing the presence and
potential impacts of waterborne environmental contaminants. Chemosphere. 54,
695-705.
Petty, J.D., Huckins, J.N., and Zajicek, J.L., 1993. Application of sempermeablemembrane devices (SPMDs) as passive air samplers. Chemosphere. 27, 1609-1624.
Pickering, R.W., 1999. A toxicological review of polycyclic aromatic hydrocarbons.
Journal of Toxicology-Cutaneous and Ocular Toxicology. 18, 101-135.
Prest, H.F., Jacobsen, L.A., and Huckins, J.N., 1995. Passive sampling of water and
coastal air via semipermeable-membrane devices. Chemosphere. 30, 1351-1361.
Prest, H.F., Jarman, W.M., Burns, S.A., Weismuller, T., Martin, M., and Huckins,
J.N., 1992. Passive water sampling via semipermfable membrane devices (SPMDs) in
concert with bivalves in the Sacramento/San Joaquin River Delta. Chemosphere. 25,
1811-1823.
Ramdahl, T., 1983. Retene – a molecular marker of wood combustion in ambient air.
Nature. 306, 580-582.
Rantalainen, A.L., Cretney, W.J., and Ikonomou, M.G., 2000. Uptake rates of
semipermeable membrane devices (SPMDs) for PCDDs, PCDFs and PCBs in water
and sediment. Chemosphere. 40, 147-158.
Rastall, A.C., Neziri, A., Vukovic, Z., Jung, C., Mijovic, S., Hollert, H., Nikcevic, S.,
and Erdinger, L., 2004. The identification of readily bioavailable pollutants in Lake
Shkodra/Skadar using semipermeable membrane devices (SPMDs), bioassays and
chemical analysis. Environmental Science and Pollution Research. 11, 240-253.
Sabaliunas, D., Lazutka, J.R., and Sabaliuniene, I., 2000. Acute toxicity and
genotoxicity of aquatic hydrophobic pollutants sampled with semipermeable
membrane devices. Environmental Pollution. 109, 251-265.
Shaw, M., Tibbetts, I.R., and Müller, J.F., 2004. Monitoring PAHs in the Brisbane
River and Moreton Bay, Australia, using semipermeable membrane devices and
90
References
EROD activity in yellowfin bream, Acanthopagrus australis. Chemosphere. 56, 237246.
Shoeib, M., and Harner, T., 2002. Characterization and comparison of three passive
air samplers for persistent organic pollutants. Environmental Science & Technology.
36, 4142-4151.
Simcik, M.F., Zhang, H.X., Eisenreich, S.J., and Franz, T.P., 1997. Urban
contamination of the Chicago coastal Lake Michigan atmosphere by PCBs and
PAHs during AEOLOS. Environmental Science & Technology. 31, 2141-2147.
Simonich, S.L., and Hites, R.A., 1995. Global Distribution of Persistent
Organochlorine Compounds. Science. 269, 1851-1854.
Smith, K.E., and Jones, K.C., 2000. Particles and vegetation: implications for the
transfer of particle-bound organic contaminants to vegetation. Science of the Total
Environment. 246, 207-236.
Smith, K.E., Thomas, G.O., and Jones, K.C., 2001. Seasonal and species differences
in the air-pasture transfer of PAHs. Environmental Science & Technology. 35, 21562165.
Stevenson, K., Bush, T., and Mooney, D., 2001. Five years of nitrogen dioxide
measurement with diffusion tube samplers at over 1000 sites in the UK.
Atmospheric Environment. 35, 281-287.
Strandberg, B., Wågman, N., Bergqvist, P-A., Haglund, P., and Rappe, C., 1997.
Semipermeable membrane devices as passive samplers to determine organochlorine
pollutants in compost. Environmental Science & Technology. 31, 2960-2965.
Sunyer, J., Ballester, F., Le Tertre, A., Atkinson, R., Ayres, J.G., Forastiere, F.,
Forsberg, B., Vonk, J.M., Bisanti, L., Tenias, J.M., Medina, S., Schwartz, J., and
Katsouyvanni, K., 2003. The association of daily sulfur dioxide air pollution levels
with hospital admissions for cardiovascular diseases in Europe (The Aphea-II study).
European Heart Journal. 24, 752-760.
Sällsten, G., Björklund, J., Johansson, O., Melin, J., Lindahl, R., Loh, C., Östman, C.,
and Barregård, L., 2001. Miljöövervakningsprojekt: Cancerframkallande ämnen i
tätortsluft - personlig exponering, individrelaterade stationära mätningar i Göteborg
2000. Report for national environmental monitoring. Swedish Environmental
Protection Agency, Stockholm, Sweden.
91
References
Södergren, A., 1987. Solvent-Filled Dialysis Membranes Simulate Uptake of
Pollutants by Aquatic Organisms. Environmental Science & Technology. 21, 855859.
Tasdemir, Y., Vardar, N., Odabasi, M., and Holsen, T.M., 2004. Concentrations and
gas/particle partitioning of PCBs in Chicago. Environmental Pollution. 131, 35-44.
Tokiwa, H., Sera, N., Nakashima, A., Nakashima, K., Nakanishi, Y., and Shigematu,
N., 1994. Mutagenic and Carcinogenic Significance and the Possible Induction of
Lung-Cancer by Nitro Aromatic-Hydrocarbons in Particulate Pollutants.
Environmental Health Perspectives. 102, 107-110.
Tremolada, P., Burnett, V., Calamari, D., and Jones, K.C., 1996. Spatial distribution
of PAHs in the UK atmosphere using pine needles. Environmental Science &
Technology. 30, 3570-3577.
Utvik, T.I, Durell, G.S., and Johnsen, S., 1999. Determining produced water
originating polycyclic aromatic hydrocarbons in North Sea waters: Comparison of
sampling techniques. Marine Pollution Bulletin. 38, 977-989.
van den Berg, M., Birnbaum, L., Bosveld, A.T., Brunström, B., Cook, P., Feeley, M.,
Giesy, J.P., Hanberg, A., Hasegawa, R., Kennedy, S.W., Kubiak, T., Larsen, J.C., van
Leeuwen, F.X., Liem, A.K., Nolt, C., Peterson, R.E., Poellinger, L., Safe, S., Schrenk,
D., Tillitt, D., Tysklind, M., Younes, M., Waern, F., and Zacharewski, T., 1998. Toxic
equivalency factors (TEFs) for PCBs, PCDDs, PCDFs for humans and wildlife.
Environmental Health Perspectives. 106, 775-792.
Voutsa, D., and Samara, C., 1998. Dietary intake of trace elements and polycyclic
aromatic hydrocarbons via vegetables grown in an industrial Greek area. Science of
the Total Environment. 218, 203-216.
Vrana, B., and Schuurmann, G., 2002. Calibrating the uptake kinetics of
semipermeable membrane devices in water: Impact of hydrodynamics.
Environmental Science & Technology. 36, 290-296.
Rogge, W.F., Hildmann, L.M., Mazurek, M.A., and Cass, G.R., 1993. Source of fine
organic aerosol. 2. Noncatalyst and catalyst-eq.uipped automobiles and heavy-duty
diesel trucks. Environmental Science & Technology. 27, 636-651.
Wania, F., Shen, L., Lei, Y.D., Teixeira, C., and Muir, D.C., 2003. Development and
calibration of a resin-based passive sampling system for monitoring persistent
organic pollutants in the atmosphere. Environmental Science & Technology. 37,
1352-1359.
92
References
Wideqvist, U., Vesely, V., Johansson, C., Potter, A., Brorström-Lunden, E., Sjöberg,
K., and Jonsson, T., 2003. Comparison of measurement methods for benzene and
toluene. Atmospheric Environment. 37, 1963-1973.
Wold, S., 1978. Cross-validatory estimations of the number of components in factor
and principal component analysis. Technometrics. 20, 397-405.
Yeo, H.G., Choi, M., Chun, M.Y., and Sunwoo, Y., 2003. Gas/particle
concentrations and partitioning of PCBs in the atmosphere of Korea. Atmospheric
Environment. 37, 3561-3570.
Yunker, M.B., MacDonald, R.W., Vingarzan, R., Mitchell, R.H., Goyette, D., and
Sylvestre, S., 2002. PAHs in the Fraser River basin: a critical appraisal of PAH ratios
as indicators of PAH source and composition. Organic Geochemistry. 33, 489-515.
Zanobetti, A., Schwartz, J., Samoli, E., Gryparis, A., Touloumi, G., Peacock, J.,
Anderson, R.H., Le Tertre, A., Bobros, J., Celko, M., Goren, A., Forsberg, B.,
Michelozzi, P., Rabczenko, D., Hoyos, S.P., Wichmann, H.E., and Katsouyanni, K.,
2003. The temporal pattern of respiratory and heart disease mortality in response to
air pollution. Environmental Health Perspectives. 111, 1188-1193.
Zimmerman, L.R., Thurman, E.M., and Bastian, K.C., 2000. Detection of persistent
organic pollutants in the Mississippi Delta using semipermeable membrane devices.
Science of the Total Environment. 248, 169-179.
93