Methodologies to assess the fate of polar organic compounds in

Methodologies to assess the fate of polar organic compounds in
aquatic environments
Jörgen Magnér
Doctoral Thesis in Applied Environmental Science
DEPARTMENT OF APPLIED ENVIRONMENTAL SCIENCE
Stockholm University, Stockholm, Sweden
2010
1
“Every decision I make is a choice between a grievance and a miracle.
I relinquish all regrets, grievances and resentments and choose the miracle.”
Deepak Chopra (M.D.)
All previously published papers were reproduced with permission from the publisher.
Printed by Universitetsservice US-AB, Stockholm, Sweden, 2010
© Jörgen Magnér, 2010
ISBN: 978-91-7447-003-1, pp 1-25
2
ABSTRACT
Polar organic compounds (POCs) are chemicals with polar functional groups in their structure. The functional groups
make the compounds hydrophilic and less prone to partition with biota. However, the knowledge of their fate is limited
due to difficulties associated with their measurements. Although, the persistence of POCs in the environment is
generally low, they are considered to be semi-persistent compounds due to their continuous introduction to the
environment via wastewater. Studies have shown that complex mixtures of POCs of different classes may have
synergistic toxic effects on biota at environmental concentration levels. Therefore, it is important to develop analytical
methods in order to establish the occurrence and fate of POCs in aquatic environments.
In Study I, a positive correlation between the sorption of a novel poly(ethylene-co-vinyl acetate-co-carbon monoxide)
(PEVAC) material and the theoretical logarithmic dissociation partition coefficient (Log D) for seven POCs was
observed. The PEVAC material showed an enhanced sorption of the POCs compared to the silicone material. Study II,
demonstrated that the PEVAC sampler assess the freely dissolved concentration of POCs in aquatic environments. The
results showed that the PEVAC polymer is an attractive alternative to silicone for mimicing the biological uptake of
POCs in aquatic environments. Additionally, Study II showed that total extraction is appropriate for determination of
the freely dissolved concentration of uncharged POCs with Log KOW < 2.67 in natural water.
In study III, a novel bag-solid phase extraction (bag-SPE) technique was compared to a conventional SPE-technique.
Despite that the extraction efficiencies for POCs in wastewater were lower using the bag-SPE method, the two methods
showed similar detection limits due to the lower ion-suppression experienced with the bag-SPE.
In study IV the bag-SPE method was further developed with the aim of lowering the detection limits for POCs.
Detection limits (LOD) below 13 ng/L showed that the bag-SPE method was suitable for determination of POCs in
surface sea water.
3
LIST OF PUBLICATIONS
This thesis is based on the following papers and manuscript. They are referred to by their roman numerals in the text.
I.
Magnér J M, Alsberg T E, Broman D (2009)
Evaluation of poly(ethylene-co-vinyl acetate-co-carbon monoxide) and polydimethylsiloxane for equilibrium
sampling of polar organic contaminants in water.
Environmental Toxicology and Chemistry 28:1874-1880
II. Magnér J M, Alsberg T E, Broman D (2009)
The ability of a novel sorptive polymer to determine the freely dissolved fraction of polar organic compounds
in the presence of fulvic acid or sediment.
Analytical and Bioanalytical Chemistry 395:1525-1532
III. Magnér J M, Alsberg T E, Broman D (2009)
Bag-SPE – A convenient extraction method for screening of pharmaceutical residues in influent and effluent
water from sewage treatment plants.
Analytical and Bioanalytical Chemistry 395:1481-1489
IV. Magnér J M, Filipovic M, Alsberg T E (XXXX)
Development of a method for determination of pharmaceutical residues in surface sea water using bag-SPE
and UPLC-QToF.
(Manuscript)
4
CONTENTS
1
Introduction
1.1 Polar organic contaminants in aquatic environments
7
1.2 Total extraction techniques
7
1.3 Bioavailability of polar organic contaminants
8
1.4 Equilibrium sampling techniques
8
2
Aims
3
Method development
4
11
3.1 The POSE sampler
12
3.2 The bag-SPE sampler
12
3.3 The PIPE sampler
13
3.4 The PEVAC sampler
13
Results and Discussion
4.1 Study I
15
4.2 Study II
15
4.3 Study III
16
4.4 Study IV
17
4.5 Future perspectives
18
5
Conclusion
20
6
Swedish summary
21
7
Statement
22
8
Acknowledgements
23
9
References
24
5
LIST OF ABBREVIATIONS
D
Dissociation constant
DOM
Dissolved organic matter
FA
Fulvic acid
FFree
Freely dissolved fraction
GC
Gas chromatography
HS
Humic substances
KAbs/W
Absorbent/water partitioning coefficient
KFA/W
Fulvic acid/water partitioning coefficient
KOC
Organic carbon/water partitioning coefficient
KOW
Octanol/water partitioning coefficient
KSed/W
Sediment/water partitioning coefficient
KTOC
Total organic carbon/water partitioning coefficient
LC
Liquid chromatography
LOD
Limit of detection
Log
Logarithmic
LOQ
Limit of quantification
MeOH
Methanol
MS
Mass spectrometry
PA
Polyacrylate
PDMS
Polydimethylsiloxane
PEVAC
Poly(ethylene-co-vinyl acetate-co-carbon monoxide)
POCs
Polar organic compounds
POM
Particulate organic matter
PS-DVB
Polystyrene-divinylbensen
SPE
Solid-phase extraction
SPME
Solid-phase micro extraction
STP
Sewage treatment plant
Tg
Glass transition temperature
6
1 INTRODUCTION
1.1 Polar organic contaminants in aquatic environments
Industrial chemicals, pesticides, pharmaceuticals and personal care products etc. are all known constituents in sewage
water from households and industries [1]. Many of these anthropogenic compounds are polar, non-volatile, and poorly
bio-degradable chemicals [2], which escape sedimentation and biological treatment in sewage treatment plants (STPs).
Therefore, STP discharges are widely accepted as the main source to the overall load of polar organic compounds
(POCs) on the aquatic environment [3,4]. Although the persistence of POCs in the environment is generally low, they
are considered to be semi-persistent and even cumulative compounds due to their continuous introduction to the
environment via wastewater [5-7]. The concentrations of POCs in wastewater have proved to be several orders of
magnitudes below the acute toxic concentrations for aquatic organisms, but less is known about possible sub-acute
effects and their wide distribution in aquatic environments have led to concern among scientists and legislative
authorities [7-9]. Recent studies have shown that complex mixtures of POCs may have synergistic toxic effects on biota
at environmental concentration levels [10-15]. Therefore, it is highly important to develop analytical methods for multiresidue screening of POCs to establish their occurrence, behaviour and fate in aquatic environments [4,16].
1.2 Total extraction techniques
Solid-phase extraction (SPE) is currently the most widely used sampling method when measuring the residues of a wide
variety of POCs from wastewater [16-19]. One of the most commonly used polymeric sorbent in SPE-cartridges, is
Oasis HLB, due to its high efficiency in retaining a diversity of POCs [3,20,21]. However, studies have shown that the
hydrophilic characteristics of Oasis HLB in combination with sampling of complex water samples will result in a low
recovery and a high degree of ion-suppression due to co-extraction of humic substances (HS) present in natural water
and wastewater [22,23]. Previous studies have shown that interferences from HS could be lowered by using mixedmode sorbents like Oasis MCX (cationic-exchanger) or Oasis MAX (anionic-exchanger)[24,25]. However, the
drawback with the ion-exchange approach is that the analysis of one water sample typically results in several different
fractions which need to be analyzed for the determination of acidic, basic and neutral pharmaceuticals [26]. Another
approach to decrease interferences from HS is to adjust the pH of the water sample to 7, which results in a
deprotonation of the HS making them negatively charged and, thus, less retained by the sorbent (Oasis HLB) than at
lower pH [27,28]. Other sorbents, without hydrophilic functional groups, e.g. polystyrene-divinylbenzene (PS-DVB)
[29], have been shown to more effectively avoid the extraction of HS at pH 7 [30].
Another problem with SPE sorbents, when applied to complex matrices, is the risk of saturating the fixed number of
surfacial bonding sites [31]. The saturation of the surface leads to competition between analytes for the limited number
of bonding sites resulting in non-linear sorption isotherms [32], which complicate the quantification of the analytes in a
sample. This is usually not a problem in conventional SPE since the capacity of the resin is dimensioned so that the
number of bonding sites significantly exceeds the total number of sorbates in the sample. Instead, the relatively polar
nature of the SPE material in combination with an overcapacity in the number of bonding sites can result in matrix
effects, such as ion suppression, due to higher enrichment of HS [18,33].
However, analysis of several types of POCs simultaneous with different physical and chemical properties generally
leads to compromises in the selection of experimental conditions, which often results in suboptimal conditions for some
analytes. Nevertheless, the development of general multi-residue analytical methods is important as it can simplify the
preparation of samples in large investigations and routine analysis of large sample sets.
7
1.3 Bioavailability of polar organic contaminants
Most organic compounds tend to bind to particulate organic matter (POM) and dissolved organic matter (DOM) present
in environmental water, usually make them less toxic to the biota [34-36]. However, the assumption that the presence of
DOM and POM in water only affects the freely dissolved concentration of compounds with octanol/water partition
coefficient (KOW) above 105 [37], and the lack of appropriate absorptive materials for sampling of compounds with KOW
less than 103 [38] have limited the knowledge regarding the fate of POCs in aquatic environments. However, in contrast
to compounds influenced mainly by hydrophobic partitioning to DOM and POM, functional groups in the structure of
POCs can bind through cohesive energy densities or through ionic interactions to functional groups of DOM and POM,
resulting in decreased freely dissolved concentrations [39-42]. Therefore, it is essential to implement bioavailability
parameters into fate models in order to be able to perform appropriate risk assessments of anthropogenic POCs present
in aquatic environments.
The distribution of a contaminant in an aquatic system can be divided into a freely dissolved fraction, a reversibly
bound fraction (associated with DOM or POM), and an irreversibly bound fraction (associated with POM) (Fig. 1)
[43,44]. The irreversibly bound fraction of a chemical will not become available for biota under any environmental
condition, and is thus less interesting in risk assessments. The reversibly bound fraction of a chemical can become
available for the biota under certain environmental conditions. The freely dissolved concentration is the fraction of an
analyte available for partitioning with biota under the given conditions and is thus denoted the bioavailable fraction.
The reversibly bound fraction together with the freely dissolved fraction of an analyte is defined as the bioaccessible
concentration. To describe the distribution of an analyte between water and DOM as a partitioning is somewhat
misconceptual, since true partitioning involves the distribution between immiscible phases, and DOM in water is
somewhere in between. Nevertheless, in lack of a more proper term, the distribution of contaminants associated with
DOM in water will here be referred to as partitioning.
The fate of anthropogenic POCs in aquatic environments has long been evaluated from SPE experiments, performed in
laboratory environments [18,45,46]. However, conventional SPE methods are designed to quantify the total amount of
an analyte in water samples, which has two major drawbacks regarding the assessment of the freely dissolved fraction
of POCs. First, the exhaustive extraction approach may disturb the initial distribution of the analytes between water,
DOM, and POM in the sample [31]. Secondly, a polar resin based sorbent, which is necessary for efficient recovery of
POCs, will also extract DOM [18], making it impossible to distinguish the DOM-bound fraction of an analyte from the
freely dissolved fraction.
1.4 Equilibrium sampling techniques
Conventional SPE techniques typically utilize an adsorptive resin consisting of a solid porous polymer with a large
surface area (10-1500m2/g). The POCs are mainly retained by specific interactions with functional groups at the surface
of the adsorbent [29]. Although the use of adsorptive polymers with specific interactions is to be preferred in certain
cases [33], there is always the risk of saturating the fixed number of surfacial bonding sites when applied to a complex
sample matrix [31].
The problem with saturation of bonding sites emerges when an adsorptive material is used as acceptor phase in passive
equilibrium sampling devices in situ, where the amount of POCs is unlimited. The limited number of bonding sites on
the surface lead to a competition between analytes and matrix components resulting in saturation and non-linear
concentration isotherms which complicates the quantification of the analytes [32]. However, passive samplers based on
adsorptive materials for in situ sampling of POCs have been developed. Such samplers, typically utilize restricted
diffusion and are operated in kinetic mode as time-integrative samplers, where the enriched amount of a POC is
8
expressed as a function of exposure time, to which the concentration of the POC in the surrounding matrix is linearly
related [47]. Another approach to minimize saturation is to utilize an absorptive instead of an adsorptive material as the
acceptor phase for POCs in in-situ passive equilibrium sampling devices. For a material to be used as an absorptive
sampler it is required that its glass transition temperature (T g) is below the temperature of employment. This is needed
in order to allow for free mobility of the polymer chains within the material. Thus, the acceptor phase operates as a
homogenous, non-porous liquid in which the analytes are retained by dissolution rather than by specific interactions
with the surface of the polymer [7,33,48,49]. This feature allows for the absorptive material to equilibrate with the
surrounding medium without reaching saturation [31]. However, it is not an absolute requisite that a sampler
equilibrates with the surrounding medium for estimation of the bioavailable concentration, but equilibrium based
techniques mimic biological uptake in a more straightforward manner than other techniques [31,43]. For example, the
addition of HS to water samples has not only proved to decrease the freely dissolved concentration of chemicals, HS
can also influence the kinetics, i.e. decrease the equilibrium time [39,50]. The effect of DOM on kinetics identifies an
advantage of equilibrium based sampling over sampling based on uptake rate, since variations in DOM concentrations
in natural waters will only affect the uptake rate, and not the equilibrium distribution between the water and the
acceptor phase.
Quantification with equilibrium sampling techniques requires that the absorbent/water partitioning coefficient (KAbs/W)
of the compound of interest is known. Furthermore, the initial equilibrium condition in the sample should be
maintained, to ensure that sample depletion is negligible [51].
In 1989 Belardi et al [52] presented the solid-phase micro extraction technique (SPME), by which a thin layer of
polydimethylsiloxane (PDMS) coated on a fused silica fiber acts as the acceptor phase. Since then, PDMS has been the
most widespread absorbent phase in equilibrium sampling devices, and used for quantitation of a variety of
contaminants in environmental applications. SPME has many advantages when assessing the bioavailable fraction of
contaminants in water. First, the low volume of acceptor phase makes negligible depletion relatively easy to achieve,
even with small sample volumes. Second, the thin absorptive film equilibrates rapidly with the surrounding media, often
within hours. The weakness of the SPME fiber is the small volume of acceptor phase which limits the amounts of a
chemical that can be enriched on the fiber, resulting in poor detection limits [53]. Another problem with the SPME
technique is the relatively apolar PDMS coating which has a limited enrichment capacity for POCs having
octanol/water partitioning coefficients (log KOW) less than 2 [54,55]. To overcome the disadvantage of poor enrichment
of more hydrophilic POCs on PDMS, polybutylacrylate (PA) was introduced as an absorptive coating of the SPME
fiber. Valor et al 2001 [55], evaluated SPME fibers coated with PA versus those coated with PDMS with regard to their
ability to enrich polar analytes. An increase in enrichment of 5-20 times was established for the selected POCs on the
PA fiber compared to the PDMS fiber. However, the SPME extraction technique rely on thermal desorption of volatile
compounds in gas chromatographic systems (GC), and many of the POCs are semi- to non-volatile substances that
preferably are separated in liquid chromatography (LC), where desorption is performed with solvents [56]. Baltussen et
al 1998 [57], demonstrated that desorption of PA materials using organic solvents proved to be a problem, because of
the large amounts of co-extracted polymer residues causing severe contamination of the ion-source of the mass
spectrometer.
9
Figure 1. Describes the distribution of contaminants in natural water
and how they can be determined using different extraction approaches.
a) Exhaustive extraction using strong solvents to establish the total
concentration. b) Gradient extraction using weak solvents to establish the
bioaccessible concentration. c) Equilibrium sampling techniques to estimate
the bioavailable fraction.
10
2 Aims
2.1 Study I
The aim of the study was to develop a passive absorptive equilibrium sampler that would enable the determination of
the concentrations of polar organic compounds (POCs) in water with higher efficiency than existing techniques.
2.2 Study II
The aim of the study was to find out if a novel absorptive materials, poly(ethylene-co-vinyl acetate-co-carbon
monoxide) (PEVAC), had the ability to distinguish the freely dissolved fraction of seven polar organic contaminants
(POCs) from the fraction bound to fulvic acid (FA) or sediment.
2.3 Study III
The aim was to develop a sampling technique for screening of pharmaceutical residues in wastewater, which offered
increased sample throughput with maintained sensitivity, precision and reproducibility compared to conventional SPE
methods.
2.4 Study IV
The aim of the study was to develop a method for determination of trace levels (ng/L) of pharmaceutical residues in
surface water that was time-saving, in terms of sample handling, compared to conventional SPE methods.
11
3 METHOD development
3.1 The POSE sampler
With the ambition to create a sampler that could monitor daily fluctuations of the freely dissolved concentration of
POCs in aquatic environments, the polar organic size-exclusion (POSE) sampler was developed (Fig. 2a). The sampler
consisted of two pieces of polyamide sheets with a thickness of 12 µm. The two parts of the plastic foils were placed on
top of each other and three of the four seams were welded with a heat-sealer. The two attached sheets formed a bag with
an opening in one end. The bag was filled with 100 mg Oasis HLB adsorbent (poly[N-vinylpyrrolidone-codivinylbenzene]) and thereafter the opening was sealed. The sampler was immersed into the sample and left
equilibrating with its surrounding under gentle stirring.
The idea behind the developed sampler was to create a size-exclusion barrier with a pore size of <1.0 nm, which would
prevent POCs bound to DOM from reaching the sorbent, while the freely dissolved fraction of POCs could pass the
barrier unhindered and be enriched on the sorbent. The polyamide foil was used as a barrier owing to its relatively polar
surface which attracts POCs. Since the polyamide foil forms a diffusive layer the sampler was operated in kinetic mode
as a time-integrative sampler.
The problem with the POSE sampler was that, despite a narrow pore size of <1.0 nm, the polyamide foil did not prevent
humic substances (HS) from accumulating on the sorbent, making it impossible to distinguish the DOM-bound fraction
of the analytes from the freely dissolved fraction. Another problem with the restricted access approach was that the
polyamide barrier decreased the diffusion rate so that it took more than a week to reach detectable levels of the
accumulated analytes, which is unacceptable considering the ambition of monitoring daily fluctuations of contaminants
in aquatic environments.
3.2 The bag-SPE sampler
In an attempt to overcome the problem with co-extraction of HS experienced with the POSE sampler a new approach
was investigated using PS-DVB as an adsorbent which, according to a previous study, exclude the extraction of HS at
pH 7 [30]. The partitioning involving the phenyl groups at the surface of the PS-DVB adsorbent and the adjacent
contaminants is characterized by weak van der Waals, dipole-dipole and induced dipole interactions, allowing the
sampler to equilibrate with its surrounding.
The developed bag-solid phase extraction (bag-SPE) sampler (Fig. 2b) consisted of two pieces of woven polyester
fabrics. The two parts of the fabrics were placed on top of each other and three of the four seams were welded with a
heat-sealer. The two attached fabrics formed a bag with an opening in one end. The bag was filled with 20 mg XAD-2
resin (polystyrene-divinylbenzene [PS-DVB]) and thereafter the opening was sealed. The sampler was immersed into
the sample and left equilibrating with its surrounding under gentle stirring.
The idea of using woven polyester fabrics (pore size of <120 µm) to enclose the PS-DVB resin (particle size 297- 840
µm) was to enable fast equilibrium to take place by eliminating diffusion through a polymer membrane.
The results showed low enrichment of HS and relatively high enrichment of the POCs, making the method suitable for
differentiating the DOM-bound fraction of an analyte from the freely dissolved fraction. However, the porous structure
(pore size <9 nm) of the PS-DVB material resulted in restricted migration of the analytes, which contributed to an
unacceptably slow equilibrium time of <7 days.
Although the sampler proved to be insufficient as a fast equilibrium sampling technique the simple handling of samples
compared to conventional SPE techniques made the method suitable for total extraction applications in small volumes
12
(10 to 100 mL).
In studies III and IV the bag-SPE sampler was evaluated as an exhaustive extraction technique for sampling of POCs in
wastewater from a STP and in surface sea water. Further information regarding the experimental set-up of the bag-SPE
method is presented in the individual studies in papers III and IV.
3.3 The PIPE sampler
During the development of the polar interface passive equilibrium (PIPE) sampler, (Fig. 2c) the idea of using phenyl
groups at the surface of the adsorbent for the partitioning of contaminants came up. However, instead of using the
porous PS-DVB as adsorbent the acceptor phase of the PIPE sampler was based on a polystyrene (PS) test tube. The
surface of the test tube was etched with acetone, which increased the efficient surface area up to 40 times, thus
enhancing the partitioning of more hydrophilic compounds. The sampler was immersed into the sample and left
equilibrating with its surrounding under gentle stirring.
The idea of using PS was to enable fast equilibration to take place by eliminating restricted migration through pores,
experienced with the bag-SPE sampler.
The result showed an equilibrium time of <24 hours with a relatively high enrichment for the selected analytes and low
enrichment of HS. Another benefit with the PIPE sampler was that it could be used with small sample volumes (10-20
ml) and still maintain negligible depletion, due to the relatively small surface area of the sampler. The method seemed
suitable for monitoring of daily fluctuations of the freely dissolved concentration of POCs in aquatic environments.
However, a problem appeared when the PIPE sampler was applied to complex sample matrices, where the amount of
POCs is unlimited. The limited number of bonding sites on the surface of the sampler leads to competition between
analytes and matrix components resulting in saturation and non-linear concentration isotherms, which complicates the
quantification of the analytes
3.4 The PEVAC sampler
To minimize saturation experienced with adsorptive materials a new approach was investigated using an absorptive
material, in which the analytes are retained by dissolution rather than by specific interactions with the surface of the
polymer. This feature allows for the absorptive material to equilibrate with the surrounding medium without reaching
saturation. Poly(ethylene-co-vinyl acetate-co-carbon monoxide) [PEVAC] was the absorptive material used in the
experiments, in the format of a polymer film (Fig. 2d). The film was made from PEVAC beads melted in boiling toluene
and poured onto aluminium foil. The hot melt was left for 24 h at room-temperature to allow toluene residues to
evaporate from the solid PEVAC film. The film, with a variation in thickness of 100-200 µm, was cut into pieces and
thereafter immersed into the sample and left equilibrating with its surrounding under gentle stirring.
The results of the PEVAC evaluation will be further discussed in the next chapter, and detailed information regarding
the experimental set-up of the method is presented in the individual studies in papers I and II.
13
Figure 2. Demonstrates the four sampler evaluated during the method
development. a) The polar organic size-exclusion (POSE) sampler. b) The
bag-solid phase extraction (bag-SPE) sampler. c) The polar interface
passive equilibrium (PIPE) sampler. d) The poly(ethylene-co-vinyl acetateco-carbon monoxide) [PEVAC] sampler.
14
4 RESULTS AND DISCUSSION
4.1 Study I
Exposing the PEVAC film to an aqueous fluoranthene solution and then illuminating the cross section in a fluorescence
microscope, illustrated that fluoranthene was distributed throughout the polymer. The result demonstrated that PEVAC
fulfilled the criteria of an absorptive material, in contrast to an adsorptive material, where the analyte would have been
located at the surface of the polymer [58].
However, fluoranthene is not the ideal model substance for evaluation of specific adsorptive interactions between polar
functional groups of POCs and the PEVAC polymer, due to the lack of polar functional groups in the fluoranthene
structure. Nevertheless, the absorptive interaction between the selected POCs and the PEVAC polymer was further
verified by the relatively small variation of the established absorbent/water partitioning coefficient (KAbs/W). The
variation was usually less than 10% for the selected compounds at neutral pH, which was small compared to the
variation of 50% in thickness of the plastic film. Should the PEVAC polymer possess adsorptive rather then absorptive
properties the differences in film thickness from 100 to 200µm would give a variation of the KAbs/W of 50% or more, due
to a 50% reduction in surface area when going from a 100µm thick PEVAC film to a 200µm thick PEVAC film. In
absorptive partitioning it is the volume of the sampler that matter for the KAbs/W, rather than the size of the surface area
[32].
Furthermore, in the present study the performance of the PEVAC polymer was validated for determination of seven
POCs in water, using a conventional PDMS method as reference [59]. The seven POCs were selected as model
substances due to their relatively even distribution along the logarithmic octanol/water partitioning coefficient (Log
KOW) scale, from Log KOW 0.2 to 4.77. Hence, it was anticipated that they would represent a majority of the POCs of
interest, as far as hydrophobic properties are concerned.
The experiment showed that six of the seven selected POCs reached thermodynamic equilibrium within four days in
both the PEVAC material and in the PDMS material.
The experimentally obtained logarithmic absorbent/water partitioning coefficient (Log KAbs/W) for six of the seven
selected compounds, in the two polymers, correlated within one order of magnitude with the theoretical logarithmic
dissociation partition coefficient (Log D). Metoprolol, however, showed a Log KAbs/W which slightly exceeded one order
of magnitudes difference from the calculated Log D. Nevertheless, a regression coefficient of R2> 0.8906 for the
correlation between the established Log KAbs/W and the calculated Log D of the compounds in the two polymers was
regarded as a relatively good agreement considering that the partitioning of compounds with polar functional groups is
not exclusively governed by hydrophobic distribution of the neutral fraction [32].
Furthermore, the PEVAC material showed an up to five times enhanced sorption for the selected compounds compared
to the silicone material. The result suggests that the PEVAC material is to be preferred over silicone for estimating
biological uptake of POCs.
4.2 Study II
In study II, a polar resin based sorbent (isolute, ENV+) was used to illustrate the problem of using an exhaustive
extraction technique for natural water. However, any sorbent with a sufficiently high density of hydrophilic elements in
its structure could have been used for the experiment. The chromatogram obtained from the SPE extract of FA rich
water showed the adsorbed FA as a broad hump extending from 1.5 to 4.0 minutes. The combined mass-spectrum of the
hump further confirmed the presence of FA in the extract (Fig. 3). Thus, detected analytes represent both the freely
15
dissolved fraction and most likely parts of the fraction originally associated with the FA, the latter being freely
dissolved in the extracting solvent, i.e. methanol. The rejection of humic substances in an extraction method is therefore
a necessity when assessing the freely dissolved concentration of a compound in aquatic environments. A combined
mass-spectrum over the same retention time interval as the hump in the SPE chromatogram was also constructed from
the chromatogram of the PEVAC sampler extract. This mass-spectrum did not show the presence of FA, which
indicates that measurements performed with the PEVAC sampler will represent the truly dissolved concentration of
contaminants in FA rich water, given that negligible depletion is achieved.
Additionally, the PEVAC sampler was used for extraction of a series of water samples, pre-spiked with seven POCs,
where the concentration of fulvic acid (FA) or suspended sediment were gradually increased in the water phase. The
results showed that the freely dissolved fraction of two of the seven selected POCs was affected by the presence of FA,
and also that the freely dissolved fraction of six of the seven POCs was affected by the presence of suspended sediment.
The FA used in the present investigation, contained carboxyl groups that would become deprotonated at pH >3.79,
which means that the carboxyl groups were negatively charged in these experiments, which were performed at pH 7.
Since metoprolol is the only compound of the investigated seven substances having a pKa sufficiently high to make it
mainly protonated and positively charged at pH 7, ionic bonding between metoprolol and FA is a likely explanation for
the drastic decrease in the freely dissolved concentration of metoprolol [40]. For the other analytes, all having Log
KOW< 5, a decrease in the freely dissolved concentration based on just hydrophobic interaction with DOM is not likely
[37], which explains why no decrease is seen for the majority of the compounds as the FA concentration was increased.
In the sediment experiment, the experimentally determined logarithmic total organic carbon partitioning coefficient
(Log KTOC) for six of the selected POCs were plotted versus their model predicted logarithmic organic carbon
partitioning coefficient (Log KOC). The model used was based on hydrophobicity [60] and demonstrated that it was
possible to predict Log KOC for compounds with Log D > 2.67 with high accuracy (R2 = 0.9887). However, strong
interactions with functional groups of the organic matter seemed to dominate the partitioning for imidacloprid,
carbendazim and metoprolol, having Log D < 1.47. Titration of the sediment with hydrochloride (HCl) and sodium
hydroxide (NaOH) showed that the carboxyl groups become deprotonated at pH > 4.2, and, as mentioned before,
because metoprolol is positively charged under the given experimental conditions (pH 7), ionic bonding to sediment
particles is a likely explanation for the drastic decrease in the freely dissolved concentration of metoprolol.
4.3 Study III
In this study the performance of the developed bag-SPE method was validated for determination of thirteen
pharmaceuticals in STP influent and effluent water, using a conventional Oasis HLB method as reference [19]. The
thirteen pharmaceuticals were selected as model substances based on their frequent occurrence in wastewater [61], and
their relatively even distribution along the Log KOW scale, from Log KOW -0.07 to 4.4.
The study showed that the majority of the selected pharmaceuticals reached distribution equilibrium between the bagSPE sampler and water within 4 hours, hence 4 hours was used as the extraction time in the following experiments.
The recoveries of the selected pharmaceuticals from the wastewater using the bag-SPE sampler ranged from 20.7 to
58.2 %, while the recoveries from the SPE-column ranged from 35.2 to 80.0 %. The results are not surprising
considering that the Oasis HLB resin in the SPE-column has been proved to have higher affinity for compounds with
polar functional groups than the XAD-2 resin, enclosed in the bag-SPE sampler [20,62]. However, the detection limits
(LOD) of the pharmaceuticals with the bag-SPE sampler (10-100 ng/L) were within the same range as the detection
limits for the compounds obtained with the SPE-column (5-100 ng/L). This is probably due to polar matrix components,
e.g. humic substances (HS), which have a higher affinity for the more hydrophilic Oasis HLB resin than for the XAD-2
16
resin, likely resulting in a higher amount of HS in the SPE sample extract. Higher concentrations of HS in an extract
would increase the ion-suppression and, thus, even out the differences in extraction efficiencies [22,27].
The dynamic range using the two extraction techniques showed high consistency with regression coefficients (R2) better
than 0.9819.
Inter-day variations determined from the analysis of both influent and effluent waters with both techniques, were lower
than 15.1 %, which was considered acceptable.
The measurements of the concentrations of pharmaceuticals in wastewater carried out on the bag-SPE sampler and on
the SPE-column, revealed highly similar results. The results showed that the concentrations of the majority of the
selected pharmaceuticals decreased as a result of the treatment in the STP.
However, the concentrations of
hydrochlorothiazide, oxazepam and carbamazepine were unaffected or even higher in the effluent water than in the
influent water. In a previous study, increased concentrations of pharmaceuticals in effluent water was explained by the
presence of conjugated compounds in the influent water, which during the treatment in the STP, can undergo
transformation into the original compounds [62]. However, since such conjugates were not included in the experiments,
no conclusions can be made regarding the fate of conjugates throughout the STP.
4.4 Study IV
In study IV the bag-SPE sampler was used to determine the concentrations of ten pharmaceuticals in surface water, The
ten pharmaceuticals were selected as model substances based on their occurrence in effluent wastewater [61], and their
relatively even distribution along the Log KOW scale, from Log KOW -0.13 to 4.39.
The study showed that the majority of the selected pharmaceuticals reached distribution equilibrium between the water
and the bag-SPE sampler within 8 hours, hence 8 hours was subsequently used as the extraction time in the experiments
to follow.
The extraction efficiency of the selected pharmaceuticals with the bag-SPE sampler ranged from 10.6 to 64.5 %, with
relative standard deviations (RSD) of < 16.4 %.
The linear concentration ranges showed high consistency with regression coefficients (R2) better than 0.9801.
The inter-day variations of the six identical sample aliquots, sampled on six different days, was less than 17.7 % for the
selected pharmaceuticals, which was considered acceptable.
Finally, with a limit of detection (LOD) of the ten analytes in the bag-SPE extract below 13 ng/L, it was demonstrated
that the method was suitable for detection of trace levels (ng/L) of pharmaceuticals in natural sea waters.
In the present study, four of the selected ten pharmaceuticals (caffeine, metoprolol, oxazepam and carbamazepine)
showed concentrations higher than the detection levels in the surface water samples from the central bay of Stockholm.
The eight sampling sites in the Stockholm area showed similar concentration levels of respective compound, which
likely have to do with the continuous introduction of pharmaceutical-residues from the two STP effluents to the
relatively closed bay.
In accordance with the samples from the Stockholm bay, the water samples collected along the coastal gradient
contained only detectable levels of caffeine, metoprolol, oxazepam and carbamazepine.
The four pharmaceuticals revealed different migration pattern throughout the gradient. For example, the concentration
of metoprolol decreased more rapidly than carbamazepine with increased distance from the STP effluent, despite that
the reported degradation times for the two compounds in water are similar [63]. One explanation could be that
metoprolol, in contrast to carbamazepine, is mainly protonated and positively charged at natural pH, and will therefore
associate with negatively charged carboxyl groups in DOM and POM present in environmental waters [40,64]. The
interaction with DOM and POM will result in a faster elimination of metoprolol from the water column, compared to
17
the neutral carbamazepine, due to sedimentation.
Caffeine showed similar concentration levels throughout the gradient. The result is confusing since caffeine is more
easily degraded compared to metoprolol and carbamazepine [63], and similar measurements have reported a decrease in
caffeine concentration with an increased distance from the STP [65]. One possible explanation to the even distribution
of caffeine along the gradient could be that private homes and summer cottages along the coast in Sweden have their
own septic tank system for their wastewater, with a weeping bed which will contribute to the overall concentration of
caffeine in the sea. The previous explanation could also be applied to the zero reference samples, lake Flaten, which
also showed detectable levels of caffeine.
Additionally, the migration of oxazepam along the coastal gradient was impossible to interpret because it was only
detected in the first two samples in the gradient.
4.5 Future perspectives
A future perspective would be to perform a study, where the partitioning of POCs to the PEVAC material is compared to
biotic uptake, to investigate whether a correlation is valid.
Study II, revealed that for POCs having Log D < 1.47, or for ionic compounds, strong interactions with functional
groups of the organic matter seemed to dominate the partitioning. Therefore, it would be interesting in future work to
model the fate of POCs in aquatic environment based on the PEVAC measurements.
18
a.
Positiv kontroll
11-Jan-2008
11:57:12
080111_7std_a
5.36
104.996
5.08
104.994
%
100
0
1.00
2.00
3.00
4.00
1: TOF MS ES+
349.926+351.926 0.50Da
2.20e3
8.52
104.989
5.00
6.00
7.00
6.00
7.00
8.00
9.00
10.00
11.00
1: TOF MS ES+
305.109 0.50Da
1.80e3
9.00
10.00
11.00
1: TOF MS ES+
216.102 0.50Da
1.18e3
10.34
080111_7std_a
100
4.71
305.115
%
1.48
214.995
0
1.00
2.00
3.00
4.61;289.161
5.08
104.994
4.00
5.00
8.06
104.995
8.00
080111_7std_a
%
100
0.02
173.967
3.29
216.107
1.91
1.60
214.995 2.63
214.995
214.995
0.83
155.980
0
1.00
2.00
8.66;155.978
3.00
4.00
5.00
6.00
7.00
8.00
3.00
4.00
5.00
6.00
7.00
8.00
8.88
9.12
214.992
173.965
9.79
173.966 155.979
9.00
10.00
9.00
10.00
080111_7std_a
2.91
237.108
11.00
1: TOF MS ES+
237.093 0.50Da
2.71e3
%
100
0
1.00
2.00
11.00
1: TOF MS ES+
256.06+258.06 0.50Da
2.16e3
080111_7std_a
2.25
256.067
%
100
4.47
261.152
0
1.00
2.00
3.00
4.00
8.67
155.978
5.19
111.022
5.00
6.00
7.00
8.00
9.00
080111_7std_a
%
100
070705_Env_FuvicAq_10ml
100
%%
2.95
1.00
2.00
3.00
0.51
070705_Env_FuvicAq_20ml
247 (2.625) Cm (130:328) 3.15 3.21
080111_7std_a
116.9064
100
1.71
100
192.082
0
0.40
0.90
173.967 173.968
0.73
1.001.19
b.
0
1.00
070705_Env_FuvicAq_20ml
11.00
1: TOF MS ES+
1: TOF MS
ES4.40e3
TIC
1.09e5
3.58
2.61
2.20 2.32
0
10.00
05-Jul-2007
14:36:09
268.191
0.50Da
2.01
268.196
2.00
3.00
2.00
3.00
4.00
5.005.19
4.15
4.35
3.76
5.32
4.76
6.00
5.90
5.73
7.00
6.14
8.00
7.29
9.00
10.00
8.55
6.63
4.57
9.08
173.965
2.32 2.63
2.93
5.00
6.00
7.00
8.00
9.00
10.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00
11.00
1: TOF MS ESTIC
1.32e5
9.00
10.00
11.00
1: TOF MS ESTIC
1.49e5
9.00
10.00
3.21
%
4.16
5.03 5.10 5.20
4.31
5.91
5.72
6.13
7.30
6.64
8.56
3.77
%
0.51
Time
11.00
4.00
3.58
100
11.00
1: TOF MS ES1: TOF MS ES+
1.63e5
192.077 0.50Da
4.15e3
0
1.00
121.0074
070705_Pevac_FuvicAq_a
2.00
3.00
4.00
5.00
c.
%
305.1088
243.0900
323.1240
295.1083
0.57
293.1014
199.0595
0.70
195.0317
2.32
1.47
2.02
187.07731.13
0
0
100
6.00
7.00
8.00
5.44
100
1.00
150
200
2.00
250
5.20
365.1433
5.64 5.91
4.78
6.64
6.10
3.70
3.29 3.35 379.14574.12
393.1479
3.15
7.07
7.29
7.65
8.27
8.56
419.1649
433.1710
3.00
300
4.34
350
4.00
400
5.00
450
500
6.00
550
600
7.00
650
8.00
700
750
800
850
900
Time
11.00
950
m/z
1000
Figure 3. a) Ion chromatograms of the selected POCs (from the top): Chlorpyrifos,
Diazinon, Atrazin, Carbamazepine, Imidacloprid, Metoprolol and Carbendazim. b) Total ion
chromatogram of the solid-phase extraction (SPE) extract showing the fulvic acid (FA) as a
hump extending from 1.5 to 4.0 min. c) The combined mass-spectrum of the hump confirming
the presence of FA in the SPE extract.
19
5 CONCLUSIONS
The main contribution of this work to the field of research, is that the study supported the assumption, previously
expressed in the literature [37], that the presence of dissolved organic matter (DOM) and particulate organic matter
(POM) in natural water does not affect the freely dissolved concentration of neutral compounds with logarithmic
octanol/water partition coefficient (Log KOW) below 5. However, if the objective is to determine the biotic exposure of
charged polar organic compounds (POCs) in natural water or of POCs associated with sediment, bioavailability
sampling techniques e.g. equilibrium sampling need to be applied. Because in contrast to compounds influenced mainly
by hydrophobic partitioning to DOM and POM, functional groups in the structure of POCs can bind, through cohesive
energy densities or through ionic interactions, to functional groups of DOM and POM, resulting in decreased
concentrations of freely dissolved POCs[39-42].
Additionally, the introduction of the novel poly(ethylene-co-vinyl acetate-co-carbon monoxide) [PEVAC] material in
the research enhanced the sorption of the selected POCs compared to PDMS based sampling techniques, which is a
significant contribution to the development of materials in the search for improving the detection of POCs in field
sampling techniques.
Since the study implicates that the presence of DOM and POM in natural water appears to not affect the freely
dissolved concentration of neutral POCs with Log KOW below 5, a novel total extraction technique for screening of
pharmaceutical residues in wastewater and surface water was developed. The novel bag-solid phase extraction (bagSPE) technique proved to be an attractive alternative to the more, in terms of sample handling, demanding solid phase
extraction (SPE) technique. Although the extraction efficiencies were lower with the bag-SPE sampler compared to the
SPE technique, the two methods showed similar detection limits due to the lower ion-suppression experienced with the
bag-SPE.
Hopefully, the results presented in this thesis will be useful in future studies on environmental fate and effects of
pollutants.
20
6 SVENSK SAMMANFATTNING
Polära organisk föreningar (POFar) är klasser av kemikaler som i sin struktur innehåller en eller flera polära
funktionella grupper. De funktionella grupperna gör föreningen mer hydrofil och därmed mindre benägen att fördela sig
till biota. Avsaknaden av lämpliga absorptiva material för polära föreningar med en oktanol/vatten fördelning (Log
KOW) mindre än 3 har begränsat kunskapen om POFars öde i akvatiska miljöer. Trots att nedbrytningen av POFar i
naturen är relativt hög, klassas de som semi-persistenta föreningar pga den kontinuerliga tillförseln av dem till miljön
via reningsverken. Studier har visat att POFar från olika klasser kan ge samverkande och skadliga effekter på biota
redan vid koncentrationsnivåer som är vanligt förekommande i naturen. Därför är det viktigt att ta fram analytiska
metoder för att fastställa förekomsten av POFar samt deras öde i akvatiska miljöer.
I studie I, påvisades en positiv korrelation mellan PEVAC-polymerens upptag och den teoretiska dissociations
konstanten (Log D) av sju POFar. PEVAC-provtagaren visade även ett ökat upptag av POFarna i jämförelse med en
traditionell provtagare av silikon. Studie II bevisade att PEVAC materialet endast anrikar den fritt lösta fraktionen av
POFar i vatten. Resultaten visar att PEVAC materialet är ett attraktivt alternativ till silikon, när det gäller att uppskatta
biologiskt upptag av POFar i akvatiska miljöer. Dessutom visade studie II att totalextraktion är tillräckligt för att
fastställa den biotillgängliga delen av POFar med Log KOW < 2.67 i naturliga vatten.
I studie III, utvecklades en ny bag-solid phase extraction (SPE) teknik som jämfördes med en konventionell SPE-teknik.
Trotts att bag-SPE metoden uppvisade en sämre extraktions effektivitet av de POFarna i reningsverks vattnet, var
detektionsgränserna jämförbara mellan de två metoderna pga den låga jonsuppressionen som erhölls med bag-SPE.
I studie IV utvecklades bag-SPE metoden ytterligare för att sänka detektionsgränserna för POFar. Detektionsgränser
(LOD) under 13 ng/L visade att bag-SPE metoden var lämplig för att bestämma koncentrationer av POFar i havsvatten.
21
7 STATEMENT
I, Jörgen Magnér, made the following contributions to the studies presented here in:
7.1 Study I
I performed the technical and methodological development of the sampling method and was responsible for the
instrumental analysis as well as the data analysis and evaluation of the two methods. I was the lead author of the paper.
7.2 Study II
I performed the methodological development and was responsible for the instrumental analysis as well as the data
analysis and validation of the results. I was the lead author of the paper.
7.3 Study III
I performed the technical and methodological development of the extraction method and was responsible for the
instrumental analysis as well as the data analysis and validation of the method. I was the lead author of the paper.
7.4 Study IV
I launched the idea to estimate the distribution of pharmaceuticals in natural surface sea water as well as planning the
project design. I was also active in the instrumental analysis as well as the data analysis and validation of the method. I
was the lead author of the manuscript.
22
8 ACKNOWLEDGEMENTS
First of all I would like to thank my supervisor Tomas Alsberg (Department of Applied Environmental Science,
Stockholm University, Sweden) for all the support and for always being available when I needed help during my PhDprogram. Second of all I would like to thank my supervisor Dag Broman (Department of Applied Environmental
Science, Stockholm University, Sweden) for sharing your positive way of thinking. I would also like to thank
Margaretha Adolfsson-Erici, Amelie Kierkegaard, Anne-Sofie Kärsrud, Katrin Holmström and Marko Filipovic
(Department of Applied Environmental Science, Stockholm University, Sweden) for making my days at the department
memorable. I thank Philipp Mayer (National Environmental Research Institute, University of Aarhus, Roskilde,
Denmark) and Michael McLachlan (Department of Applied Environmental Science, Stockholm University, Sweden) for
the knowledge and support regarding partitioning and equilibrium sampling techniques. I also thank Tomas Hansson
(Department of Applied Environmental Science, Stockholm University, Sweden) and Eva Eklöf (Swedish Museum of
Natural History, Department of Contaminant Research, Stockholm, Sweden) for the support regarding fluorescencemicroscopy. Finally but not least, I would like to thank my girlfriend Tania Barbagianni for making my dreams come
true.
This research was financially supported by European Union (European Commission, FP6 Contract No. 003956) “Novel
Methods for Integrated Risk Assessment of Cumulative Stressors in the Environment” (NoMiracle) and by the Swedish
research council Formas.
23
9 REFERENCES
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
Daughton C G, Ternes T A (1999) Environ Health Perspect 107:907-944
Paxéus N (2004) Water Sci Technol 50:253-260
Buchberger W W (2007) Anal Chim Acta 593:129-139
Kolpin D W, Furlong E T, Meyer M T, Thurman E M, Zaugg S D, Barber L B, Buxton H T (2002) Environ Sci
Technol 36:1202-1211
Hernando M D, Mezcua M, Fernández-Alba A R, Barceló D (2006) Talanta 69:334-342
Bendz D, Paxéus N A, Ginn T R, Loge F J (2005) J Hazard Mater 122:195-204
Muñoz I, Gómez J M, Molina-Díaz A, Huijbregts M A J, Fernández-Alba A R, García-Calvo Eloy (2008)
Chemosphere 74:37-44
Quinn B, Gagné F, Blaise C (2008) Sci Total Environ 389:306-314
Richards SM, Cole S E (2006) Ecotoxicology 15:647-656
Hernando M D, Fernández-Alba A R, Tauler R, Barceló D (2005) Talanta 65:358-366
Fraker S L, Smith G R (2004) Environ Toxicol 19:250-256
Pomati F, Castiglioni S, Zuccato E, Fanelli R, Vigetti D, Rossetti C, Calamari D (2006) Environ Sci Technol
40:2442-2447
Flaherty C M, Dodson S I (2005) Chemosphere 61:200-207
Marinovich M, Ghilardi F, Galli C L (1995) Toxicology 108:201-206
Jaeger J W, Carlson I H, Porter W P (1999) Toxicol Ind Health 15:133-151
Zhang Z L, Zhou J L (2007) J Chromatogr A 1154:205-213
Petrovic M, Gros M, Barcelo D (2006) J Chromatogr A 1124:68-81
Masqué N, Marcé R M, Borrull F (1998) Trends in Analytical Chemistry 17:384-394
Gros M, Petrović M, Barceló D (2006) Talanta 70:678-690
Fontanals N, Marcé R M, Borrull F (2007) J Chromatogr A 1152:14-31
Santos J L, Aparicio I, Alonso E, Callejón M (2005) Anal Chim Acta 550:116-122
Renew J E, Huang C-H (2004) J Chromatogr A 1042:113-121
Chen H-C, Wang S-P, Ding W H (2006) J Chromatogr A 1102:135-142
Himmelsbach M, Buchberger W, Klampfl C W (2006) Electrophoresis 27:1220-1226
Kasprzyk-Hordern B, Dinsdale R M, Guwy A J (2007) J Chromatogr A 1161:132-145
Lavén M, Alsberg T, Yu Y, Adolfsson-Erici M, Sun H (2009) J Chromatogr A 1216:49-62
Gómez M J, Petrović M, Fernández-Alba A R, Barceló D (2006) J Chromatogr A 1114:224-233
Weigel S, Kallenborn R, Hühnerfuss H (2004) J Chromatogr A 1023:183-195
Filippov O A, Tikhomirova T I, Tsizin G I, Zolotov Y A (2003) J Anal Chem 58:398-422
Pichon V, Coumes C D, Chen L, Guenu S, Hennion M-C (1996) J Chromatogr A 737:25-33
Mayer P, Tolls J, Hermens J L M, Mackay D (2003) Environ Sci Technol 37:184A-191A
Schwarzenbach R P, Gschwend P M, Imboden D M (2003) Environmental Organic Chemistry. Hoboken, New
Jersey, USA
Baltussen E, Cramers C A, Sandra P J F (2002) Anal Bioanal Chem 373:3-22
Traina S J, Mcavoy D C, Versteeg D J (1996) Environ Sci Technol 30:1300-1309
Day K E (1990) Environ Toxicol Chem 10:91-101
Boa M L, Dai S G, Pantani F (1997) Environ Contam Toxicol 59:671-676
Chin Y-P, Weber W J Jr, Eadle B J (1990) Environ Sci Technol 24:837-842
Magnér J, Alsberg T, Broman D (2009) Environ Toxicol Chem 28:1874-1880
Poerschmann J, Kopinke F-D, Pawliszyn J (1997) Environ Sci Technol 31:3629-3636
Lützhøft H C H, Vaes W H J, Freidig A P, Halling-Sørensen B, Hermens J L M (2000) Environ Sci Technol
34:4989-4994
Haderlein S B, Schwarzenbach R P (1993) Environ Sci Technol 27:316-326
Kopinke F-D, Pörschmann J, Remmler M (1995) Naturwissenschaften 82:28-30
Reichenberg F, Mayer P (2006) Environ Toxicol Chem 25:1239-1245
Semple K T, Doick K J, Jones K C, Burauel P, Craven A (2004) Environ Sci Technol 38:228A-231A
Sancho J V, Pozo O J, Hernández F (2004) Analyst 129:38-44
Bagheri H, Saraji M, Barceló D (2004) Chromatographia 59:283-289
Mills G A, Vrana B, Allan I, Alvarez D A, Huckins J N, Greenwood R (2007) Anal Bioanal Chem 387:11531157
Reiter G (2002) Eur Phys J E 8:251-255
Danch A, Osoba W (2003) J Thermal Analysis and Calorimetry 72:641-650
Mayer P, Karlson U, Christensen P S, Johnsen A R, Trapp S (2005) Environ Sci Technol 39:6123-6129
Vaes W H J, Ramos E U, Verhaar H J M, Seinen W, Hermens J L M (1996) Anal Chem 68:4463-4467
Belardi R P, Pawliszyn J B (1989) Water Quality Research J Canada 24:179
Heringa M B, Hermens J L M (2003) Trends in Analytical Chemistry 22, 575-587
Blasco C, Font G, Picó Y (2002) J Chromatography A 970:201-212
24
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
Valor I, Pérez M, Cortada C, Apraiz D, Moltó J C, Font G (2001) J Sep Sci 24:39-48
Chen J, Pawliszyn J B (1995) Anal Chem 67:2530-2533
Baltussen E, David F, Sandra P, Janssen H G, Cramers C (1998) J High Resol Chromatogr 21:645-648
Mayer P, Vaes W H J, Hermens J L M (2000) Anal Chem 72:459-464
Paschke A, Brümmer J, Schüürmann G (2007) Anal Bioanal 387:1417-1421
Karickhoff S W (1981) Chemosphere 10:833-846
Carlsson C, Johansson A K, Alvan G, Bergman K, Kuhler T (2006) Sci Total Environ 364:67-87
Radjenivic J, Petrovic M, Barceló D (2007) Anal Bioanal Chem 387:1365-1377
Kim I, Yamashita N, Tanaka H (2009) J Hazard Mater 166:1134-1140
Magnér J A, Alsberg T E, Broman D (2009) Anal Bioanal Chem 395:1525-153
Weigel S, Berger U, Jensen E, Kallenborn R, Thoresen H, Hühnerfuss H (2004) Chemosphere 56:538-592
25