Chemosphere 122 (2015) 105114

Contents lists available at ScienceDirect

Chemosphere
journal homepage: www.elsevier.com/locate/chemosphere

Development and application of a simultaneous SPE-method for

polycyclic aromatic hydrocarbons (PAHs), alkylated PAHs, heterocyclic
PAHs (NSO-HET) and phenols in aqueous samples from German Rivers
and the North Sea
Anne-Kathrin Siemers, Jan Sebastian Mnz, Wolf-Ulrich Palm , Wolfgang K.L. Ruck
Leuphana University Lneburg, Institute of Sustainable and Environmental Chemistry, Scharnhorststr. 1, 21335 Lneburg, Germany

h i g h l i g h t s
 Two simultaneous extraction methods for PAHs and NSO-HETs.
 First comprehensive data on concentrations of NSO-HETs in surface waters.
 N-HETs (acridine, quinoline and methylquinolines) prevailed.

a r t i c l e

i n f o

Article history:
Received 30 June 2014
Received in revised form 31 October 2014
Accepted 8 November 2014
Available online 4 December 2014
Handling Editor: Keith Maruya
Keywords:
Extended surface water monitoring
program
PAHs
NSO-HETs
GCMS
LCMSMS

a b s t r a c t
Polycyclic aromatic hydrocarbons (PAHs), heterocyclic PAHs (NSO-HETs), alkylated PAHs and phenols are
known as the prevailing contaminants in groundwater at tar contaminated sites. Besides these local
sources, the concentrations and the distribution in particular of NSO-HETs in environmental samples,
such as rivers, have received notably less attention. To investigate their occurrence in river basins two
sensitive analytical methods for the simultaneous extraction of 86 substances including NSO-HETs,
classical EPA-PAHs, alkylated PAHs and phenols were developed: liquidliquid extraction for the whole
water phase and solid phase extraction for the dissolved water phase only. Solely GCMS or additionally
LCMSMS for fractionated basic nitrogen heterocycles (N-HETs) were used for quantication. Limits of
quantication were in the low ng L 1 range. Concentrations were determined in 29 aqueous samples
from 8 relatively large rivers located in Lower Saxony (Germany) and the North Sea. NSO-HETs had comparable or even higher sum concentrations than EPA-PAHs. N-HETs, especially acridine and quinolines
with concentrations of up to 20 ng L 1 per substance, were predominant.
2014 Elsevier Ltd. All rights reserved.

1. Introduction
PAHs are a group of persistent organic pollutants which were
extensively studied in various kinds of matrices over the last
decades. Their origin can be pyrolytic or petrogenic, so that PAHs
are used as markers for combustion processes, fuel spills or taroil contaminations. Owing to their ubiquitous distribution in the
environment and their carcinogenic activity, some PAHs have been
attracting research attention. After the US Environmental Protection Agency (EPA) classied 16 PAHs as priority pollutants in the
early 1980s many researchers focused on that list, thereby neglecting (Andersson et al., 2006) other related substances present in the
Corresponding author. Tel.: +49 4131 6772874; fax: +49 4131 6772822.
E-mail address: palm@uni.leuphana.de (W.-U. Palm).
http://dx.doi.org/10.1016/j.chemosphere.2014.11.022
0045-6535/ 2014 Elsevier Ltd. All rights reserved.

environment, such as heterocyclic PAH analogues (NSO-HETs, also

termed hetero-PAHs). NSO-HETs contain nitrogen (N-HETs), oxygen (O-HETs) or sulfur (S-HETs) atoms within their ring structure.
This leads, in particular for the N-HETs, to a signicant increase of
polarity and water solubility. For example, naphthalene as the simplest PAH possesses a water solubility of approximately 30 mg L 1
whereas the solubility of the corresponding N-HET quinoline is
6100 mg L 1 (EPI-Suite, 2010, experimental values at 25 C). Table 1
shows physicochemical properties of some of the analyzed NSOHETs.
NSO-HETs and alkylated PAHs reentered scientic focus and
discussion in the course of the investigation and remediation of
contaminated sites (Blotevogel et al., 2008; Schlanges et al.,
2008). The toxicity and bioaccumulation of some NSO-HET representatives was shown in several studies (e.g. Bleeker et al., 1999;

106

A.-K. Siemers et al. / Chemosphere 122 (2015) 105114

Feldmannova et al., 2006; Peddinghaus et al., 2012). Recently

Brinkmann et al. (2014) determined estradiol equivalence factors
for selected NSO-HETs: For acridine, xanthene, indole, 2-methylbenzofuran, 2,3-dimethylbenzofuran, dibenzofuran, dibenzothiophene, quinoline, and 6-methylquinoline, these values were
comparable to those of other xenoestrogens (e.g., alkylphenols or
bisphenol A). Although tar-oil contaminated sites are a known
source of NSO-HETs and especially coking plants were often
located adjacent to rivers, only scarce data exist on their occurrence in the aqueous phase of sea water or even rivers themselves.
Shinohara et al. (1983) determined 14 N-HETs in Japanese sea
water with concentrations of up to 55 ng L 1 of 2,4-dimethylquinoline. S-HETs in sea water samples from the Catalonian coast of
Spain were below limits of quantication (LOQ) which ranged from
3 up to 30 ng L 1 (Gimeno et al., 2002). More recent data based on
GCMS screening from Slovak rivers showed maximum concentrations of two N-HETs: indole with 84 lg L 1 and quinoline with
25 lg L 1 (Slobodnik et al., 2012). Furthermore some heterocycles
such as acridine and acridinone are known to be transformation
products of the antiepileptic drug carbamazepine (Kosjek et al.,
2009). Lpez-Serna et al. (2012) detected concentrations of up to
18 ng L 1 for both substances in the Spanish River Ebro and its
tributaries.
As mentioned above the regulation of the group of polycyclic
aromatic compounds in the environment is restricted almost
exclusively to EPA-PAHs. Within the European Union the Water
Framework Directive (WFD) was established in 2000 with the
aim to achieve a good ecological and good chemical status of surface waters in Europe by 2015 (European Commission Directive
2000/60/EC, 2000). To characterize the chemical status of a river,

Environmental Quality Standards (EQS) for priority substances

were dened, among them several EPA-PAHs. The annual average
EQS for regulated PAHs in the whole water sample range from
0.002 up to 2.4 lg L 1 (European Commission Directive 2008/
105/EC, 2008). Furthermore, the member states are requested to
identify more compounds of concern. One possible rst step is to
monitor the occurrence of potentially hazardous substances. In
the case of PAHs in the aqueous phase, NSO-HETs must be taken
into consideration due to their solubility.
In general, the analytical methods used for polycyclic aromatic
compounds consist of an extraction, a clean-up and an enrichment
step prior to gas chromatographic (GC) or liquid chromatographic
(LC) separation and determination by mass spectrometry (MS).
Widely used extraction methods are liquidliquid extraction
(LLE) or solid phase extraction (SPE). For EPA-PAHs many methods
are standardized (US EPA, 1996; US EPA, 2004; ISO, 2011).
NSO-HETs can be extracted in similar manner e.g. by LLE using
dichloromethane (DCM) (Johansen et al., 1996) as well as by SPE
using cyclohexyl sorbents (Rostad et al., 1984), polystyrene
divinylbenzene sorbents (Mundt and Hollender, 2005) or modied
silica sorbents (Jeanneau et al., 2007; Johansen et al., 1996). Many
N-HETs possess basic properties which can be used for fractionation by modication of the pH value (Steinheimer and Ondrus,
1986). Because of their high water solubility, phenols have to be
derivatized with diazomethane, pentauorobenzyl bromide or
pentauorobenzyl chloride (US EPA, 1984; Renberg, 1981) prior
to LLE. By using SPE technique, EPA Method 528 (US EPA, 2000)
leaves it open to derivatize phenols. In general, determination is
carried out by GCMS, but for N-HETs LCMSMS is highly
recommended.

Table 1
Selected NSO-HETs investigated including their CAS registry number, formulae and physicochemical properties.
Compound

CAS registry number

Chemical formula

Acridine

260-94-6

C13H9N

9(10H)-Acridinone

578-95-0

C13H9NO

Structural formula

Molar mass (g mol

Water solubility (mg L

a

log KOW
3.4b

179.22

38.4(25 C)

195.22

13.53(25 C)

2.84

129.16

6110(25 C)a

2.03b

143.19

498.5(25 C)c

2.59b

145.16

1050(20 C)a

1.26b

184.26

1.47(25 C)a

4.38b

212.31

0.25(25 C)c

5.26c

168.20

3.1(25 C)a

4.12b

O
c

N
H

Quinoline

91-22-5

C9H7N

2-Methylquinoline

91-63-4

C10H9N

2(1H)-Quinolinone

59-31-4

C9H7NO

CH3

Dibenzothiophene

132-65-0

C12H8S

4,6-Dimethyl-dibenzothiophene

1207-12-1

C14H12S

S
CH3

Dibenzofuran

132-64-9

CH3

C12H8O
O

a
b
c

EPI-Suite experimental value.

EPI-Suite experimental value (Hansch et al.).
EPI-Suite estimated value.

A.-K. Siemers et al. / Chemosphere 122 (2015) 105114

Signicant differences in polarity, pH-dependent properties of

N-HETs, the purpose to handle sample preparation without
derivatization and the simultaneousness are challenging. The aim
of this study was to develop a LLE and a SPE analytical method
for determining the classical EPA-PAHs, including regulated compounds, NSO-HETs, alkylated PAHs and phenols simultaneously
in the ng L 1 range. Aqueous samples were quantied using solely
GCMS or additionally LCMSMS. Furthermore, recovery experiments using ultrapure water, river- and sea water were performed
to investigate the inuence of different matrices. Finally, these
methods were applied to the analysis of surface water samples
from relatively large river systems in Northern Germany and the
North Sea.
2. Materials and methods
2.1. Reagents, standards, and apparatus
Solvents (acetone, acetonitrile, dichloromethane, n-hexane,
methanol, toluene) were purchased in the highest available purity
from Scharlau (Sentmenat, Spain), VWR (Prolabo, Darmstadt,
Germany) and Merck (Darmstadt, Germany), hydrochloric acid
(34%) and ammonia solution (20%) from VWR (Prolabo, Darmstadt,
Germany), sodium azide and anhydrous sodium sulfate from
Merck (Darmstadt, Germany). All analyzed 86 organic compounds
(34 heterocyclic PAHs, 17 alkylated PAHs, the 16 classical
EPA-PAHs, 14 phenols and 5 others) as well as deuterated internal
standards (IS) were purchased as reference standards and CAS
numbers and suppliers are summarized in the supplementary
materials. Deuterated internal standards were used in three different solutions (each c = 1 mg L 1): (1) IS-GC: naphthalene-D8,
anthracene-D10 and perylene-D12 in toluene, (2) IS-LC: quinoline-D7 in acetonitrile (ACN) both used as injection standards
and (3) IS-EX: phenol-D6, acenaphthene-D10 and acridine-D9 in
ACN, which was added to every sample prior to extraction. Ultrapure water (18.2 MX cm 1) was obtained from a SG Ultraclear
water purication system (Barsbttel, Germany). A Heidolph
(Schwabach, Germany) Reax 2 overhead shaker was used. Filtration was realized by using a Sartorius (Gttingen, Germany) standard lter apparatus and glass ber lters (GF-A, Whatman,
1.6 lm). Supelco (Bellefonte, USA) vacuum extraction manifolds
and Lichrolut EN cartridges (200 mg, 3 mL) from Merck (Darmstadt, Germany) were used. Tubing was made of stainless steel
(inner diameter 1 mm, length 600 mm). A Heidolph (Schwabach,
Germany) Laborota 4003 control rotary evaporator was used for
sample concentration. Prior to use, all glassware was rinsed with
organic solvents and ultrapure water and baked out at temperatures between 180 C and 300 C for at least 2 h.
2.2. Study area and sample collection
Lower Saxony in the north-western part of Germany is the
second largest among the sixteen federal states of the country.
Three big river basins (River Elbe, River Ems, River Weser) are
mainly or partly located within that state and were investigated
during the sampling campaigns. Fig. 1 shows a map of all sampling
sites. The rst campaign (Lower Saxony) was conducted in June
2009 with the primary objective to screen, if in particular
NSO-HETs are present in the aqueous phase. Samples were taken
2.5 m from the river bank (water depth < 0.5 m) directly into precleaned 1 L amber glass bottles. For conservation sodium azide
solution (c = 70 g L 1) was added. The samples were stored at
4 C and extracted by using the LLE method. The obtained data
are concentrations for the whole water phase as demanded in
the WFD.

107

The second campaign took place in October 2010 in the estuaries of the rivers Elbe and Weser with the chief aim to investigate
concentrations in the dissolved aqueous phase along two riverine
courses. On board of the research vessel Ludwig Prandtl, surface
water samples (water depth < 0.5 m) were taken in pre-cleaned
1 L amber glass bottles, ltered through glass ber lters and
acidied with hydrochloric acid to pH 12. These samples were
extracted by using the more feasible SPE method to determine
concentrations for the dissolved phase.
Major goal of the third campaign in February 2012 was to
examine the occurrence of the substances in the dissolved phase
in the North Sea. The seven selected sampling sites are part of
the Coastal Observation System for Northern and Arctic Seas
(COSYNA). On board of the research vessel Heincke, sea water
samples were taken and treated like in the second campaign
described above.
2.3. Sample preparation
Fig. 2 gives a schematic overview about the two developed
extraction methods.
Without derivatization the LLE method does not permit the
extraction of phenols from water because of their high water
solubility. Furthermore, a high sample throughput is impractical.
Therefore, a combined extraction strategy using SPE and
subsequent fractionation by pH variation on Lichrolut EN cartridges was developed. To control fractionation efciency of the
cartridges the ltered aqueous samples were spiked with 25 lL
IS-EX solution. For the enrichment of the aqueous phase using
stainless steel tubing, a ow rate was adjusted to 34 mL min 1
to load the cartridges. Overall 1 L water sample was concentrated
to 200 lL extract. An aliquote of 100 lL of the extract was prepared
for the analysis by adding 5 lL of IS. The SPE procedure separates
basic nitrogen heterocycles (2nd fraction) and possesses a suitable
clean-up prior to LCMSMS analysis. Quantication of all other target substances (1st fraction) like phenols, PAHs, S- and O-HETs was
realized by GCMS. Detailed sample preparation protocol can be
retrieved from supplementary materials.
2.4. Analytical procedure
For GCMS analysis, a Thermo Finnigan Trace 2000 gas chromatograph with an Optima 5-MS capillary column (Macherey
Nagel), coupled with a Trace DSQ quadrupole mass spectrometer
was used. MS operated in positive electron ionization mode.
LCMSMS analysis were carried out by an Agilent 1200 series Innity liquid chromatograph coupled with a 6430 triple quadrupole
mass spectrometer. A Nucleodur C18-PAH column (Macherey
Nagel) and a corresponding guard column were used. The postcolumn addition of formic acid induced an acidication of the substances in order to increase the signal intensity. MSMS operated in
positive electrospray ionization mode.
A list of all substances including CAS registry number, quantier
and qualier for GCMS, precursor ion, product ion, collision
energy and fragmentor voltage for LCMSMS and retention time,
the analytical procedure in detail as well as further sample characterizations are given in supplementary materials.
2.5. Calibration, limits of quantication, recoveries, quality assurance
and quality control (QA/QC)
For GCMS analytics nine-point calibrations were performed
with every sample batch using serially diluted working solutions
in toluene (6-1666 lg L 1). For LCMSMS analytics ten-point calibrations were performed with every sample batch using serially
diluted working solutions in ACN (21000 lg L 1) containing all

108

A.-K. Siemers et al. / Chemosphere 122 (2015) 105114

Fig. 1. Sampling sites of the three campaigns, E = River Elbe, W = River Weser, N = North Sea, boxes = big cities, next to the triangles the name of the rivers investigated in the
Lower Saxony Campaign is specied.

Fig. 2. Schematic overview of the liquidliquid and the solid-phase extraction method, DCM = dichloromethane.

N-HETs which possess basic properties. Additionally further diluted

solutions down to 0.4 lg L 1 for GCMS and down to 0.03 lg L 1 for
LCMSMS were analyzed to control the linearity of the detector.
Referring to DIN method 32645 (DIN, 1994) LOQ values of the
instruments were obtained from standard solutions and were
approximately 0.52.3 ng L 1 for EPA-PAHs, 0.42.4 ng L 1 for
alkylated PAHs, 0.20.8 ng L 1 for S-HETs, 0.69.8 ng L 1 for O-HETs

and 0.22.1 ng L 1 for basic N-HETs. LOQs are given in detail in supplementary data. Every batch of samples contained at least one laboratory blank out of ultrapure water which was treated in the same
way as environmental samples, solvent blanks which were concentrated and working solutions as quality control measures.
To examine the extraction efciency, the loss of analytes during
the whole method process and the ruggedness of the method,

A.-K. Siemers et al. / Chemosphere 122 (2015) 105114

109

Fig. 3. LCMSMS chromatogram, dashed line shows the percentage of ACN in the gradient elution.

recovery experiments with different water matrices were

performed. Method recoveries and errors for each analyte were
determined by three extractions of spiked solutions in ultrapure
water, in ltered river water and in ltered sea water. Only
samples with acridine-D9 recoveries within 10% range of the
previously determined acridine recovery of the corresponding
water type were integrated for quantication. The raw concentration data of environmental samples were checked for LOQ, corrected by laboratory blank value, recovery and enrichment factor.
Total concentrations of substance groups were calculated as the
sum of all associated analytes determined.

3. Results and discussion

3.1. Analytical methods
A Total Ion Chromatogram of a working solution in toluene and
three deuterated substances is shown in supplementary materials.
Xcalibur Software (Thermo Scientic) was used for data acquisition and analysis. Depending on retention time and qualier/quantier ion ratio, peak identication was carried out. Calibration
functions were found to be linear with typical correlation coefcients >0.995. Twelve pairs of substances could not be separated

110

A.-K. Siemers et al. / Chemosphere 122 (2015) 105114

during chromatography on the used Optima 5-MS capillary column. In the case of co-eluting constitutional isomers in ve pairs
(2-/3-methylphenol, 2,4-/2,5-dimethylphenol, 6-/7-methylquinoline, 2,6-/2,7-dimethylnaphthalene and 1,4-/2,3-dimethylnaphthalene) had to be quantied together. With increasing number of
injections the entire group of N-HETs began to tail with shifting
retention times, so that the column had to be cut regularly or even
replaced.
For LCMSMS method development individual working solutions were injected. In order to obtain good separation of substances prior to mass detection, at the beginning six different
columns were tested (see supplementary materials). The best separation was achieved under neutral conditions using a mobilephase gradient. As seen in Fig. 3 the two methylisoquinolines as
well as the isomers 3-/4-/7- and 8-methylquinoline elute in one
peak and had to be quantied together. In general the N-HETs were
baseline-separated using GC, whereas using LC some substances
could not be baseline separated. By this action, we could conrm
the results by Delhomme and Millet (2008), that GC provides a
higher efciency of the separation of N-HETs than LC.
The great advantage of LCMSMS in N-HETs analysis however,
is the high sensitivity and the peak symmetry. Especially the polar
compounds like for example 6(5H)-phenanthridinone, which
showed asymmetric peaks in form of tailing during GC analysis,
could be quantied at much lower LOQ-values.
To enhance signal intensity, a post-column addition of
10 lL min 1 formic acid (1%) to the eluent was used which induced
an acidication of the substances. For the non-alkylated two-ring
compounds quinoline and isoquinoline an increase of signals by
a factor 2025, and for the alkylated quinolines an increase by a
factor 410 was achieved. For the three-ring compounds only a
slight improvement of the signal intensity was observed, whereas
for the three keto-compounds a slight decrease was found. Due to
the positive effect on the majority of the analytes the post column
derivatization was carried out in all LCMSMS measurements.
For data acquisition and evaluation, the MassHunter WorkStation Software (Agilent Technologies) was used. Quantication was
performed by applying an internal standard method. All calibration
curves were linear and correlation coefcients were >0.995.

3.2. Sample preparation protocol

For LLE from preliminary experiments two different solvents,
toluene and n-hexane, were chosen. By the application of 20 mL
hexane in the rst and 10 mL toluene in the second extraction step,
the best overall yield for all compounds (without phenols) of
78 7% was achieved. For laboratory and separating-funnel blanks
no concentrations above LOQ were determined.
For the SPE method different sorbent materials were examined:
silica-based sorbents were not suitable to extract phenols and these
sorbents are not stable in a basic pH-range, preventing a fractionation of samples. For this spectrum divinylbenzene co-polymers
were found to be suitable. Direct application of a working solution
on cartridges in comparison to enrichment from 100 mL spiked
ultrapure water with tubings made of PTFE and stainless steel (each
n = 3) was realized to evaluate the inuence of tubing on recoveries.
On average stainless steel tubing showed up to 20% higher recoveries for the highly non-polar EPA-PAHs and was used for all extractions. Furthermore different solvent mixtures (acetone/DCM (1:2,
v/v), methanol/DCM (1:2, v/v) acetone/methanol/DCM (1:1:2, v/v/
v)) were compared for the elution of the 2nd fraction. The highest
recoveries P95% were obtained by using the mixture of acetone/
DCM. Two cartridges inserted into each other provided the
opportunity to identify any overload and substance break through.
However, no breakthrough was observed, not even in recovery
experiments. Laboratory blanks for SPE for some compounds were
found to be slightly above the LOQ: indole had blank concentrations
in the range of 25 ng L 1 excluding a possible quantication.
Furthermore this method does not allow the quantication of
phenol and methylphenols (cresols) due to assumed laboratory
contaminations.
Table 2 shows mean recoveries and relative standard deviations
(RSD) for the different compound groups as well as precisions of
the methods, calculated as RSD of the recoveries. In general the
SPE method was more practical and achieved higher recoveries
in all three water types used. However, recoveries of N-HETs were
affected by the matrix. Acridine-D9 is appropriate to assure and to
correct fractionation efciency. Hence, the fractionation possesses
a suitable clean-up prior to LCMSMS analysis.

Table 2
Compound-group recoveries und precisions for LLE and SPE, ne = not extractable, n = 3, n = 2.
Compound groups

c
d
e

SPE-method
Ultrapure water
c = 25 ng L 1
Recovery (%)

SPE-method
River water
c = 25 ng L 1
Recovery (%)

SPE-method
Sea water
c = 25 ng L 1
Recovery (%)

Mean

RSD

Mean

RSD

Mean

RSD

Mean

RSD

N-HETs
S-HETs
O-HETs
EPA-PAHs
Alkyl-PAHs
Phenols

65a
65
51
82
63
ne

26
14
17
22
14
ne

88b
98
85
85
86
80e

17
17
20
21
20
20

71c
85
84
75
89
106e

17
9
24
25
20
15

61d
88
80
83
84
92e

12
8
27
25
14
10

Compound groups

Precision (%)

N-HETs
S-HETs
O-HETs
EPA-PAHs
Alkyl-PAHs
Phenols
b

LLE-method
Ultrapure water
c = 8 ng L 1
Recovery (%)

Without
Without
Without
Without
Without

Precision (%)

Mean RSD

RSD (RSD)

17a
18
14
7
23
ne

15
14
8
8
6
ne

Mean RSD
5b
4
3
3
2
8e

Precision (%)
RSD (RSD)
4
5
2
2
1
7

2-hydroxycarbazole, phenanthridinone, 4-phenylpyridine.

2-hydroxycarbazole, indole, 9-methylacridine, 2-phenylpyridine.
9(10H)-acridinone, 2-hydroxycarbazole, indole, 9-methylacridine, 2-phenylpyridine.
2-hydroxycarbazole, indole and 9-methylacridine.
phenol and cresols.

Mean RSD
13c
4
4
7
3
2e

Precision (%)
RSD (RSD)

Mean RSD

8
2
7
10
2
1

18d
7
8
6
8
12e

RSD (RSD)
8
2
2
3
2
6

Table 3
Environmental Quality Standards (EQS) of EPA-PAHs for surface waters, insignicance thresholds of NSO-HETs for groundwater and determined concentrations in the Lower Saxony Campaign in ng L
average, MAC = maximum allowable concentration, na = not applicable, P EQS, P insignicance threshold, no threshold values available.

. AA = annual

A.-K. Siemers et al. / Chemosphere 122 (2015) 105114

111

112

A.-K. Siemers et al. / Chemosphere 122 (2015) 105114

In the low concentration regime (up to 100 lg L 1) 2-hydroxycarbazole, phenanthridinone and 4-phenylpyridine could not be
determined by GCMS due to asymmetric peak shapes. For
2-phenylpyridine increased recovery values were found, but could
not be traced back to contaminations. The N-HET 9-methylacridine
could not be quantied using SPE. Further experiments (details not
discussed here) conrm that 9-methylacridine is transformed on
the sorbent material into 9-acridine carboxaldehyde.
3.3. Environmental samples
The results are presented in a chronological order of the three
campaigns. Concentrations are summarized; all data can be seen
in detail in the supplementary materials.
3.3.1. Large Rivers in Lower Saxony (First campaign)
Comparable sum concentrations of heterocyclic and homocyclic
compounds were found in river samples in the whole water phase.
Out of the heterocyclic compounds analyzed, dibenzothiophene,
4,6-dimethylbenzothiophene, quinoline, 2-methylquinoline, acridine and carbazole were found in at least seven out of nine samples.
The sum of the 16 EPA-PAHs ranged from 13 to 140 ng L 1. Schfer
et al. (2011) provide data of maximum concentrations of more than
300 organic compounds measured between 1994 and 2004 in the
rivers of interest. Among them were EPA-PAHs (up to 200 ng L 1)
and some phenols (up to 710 lg L 1). However, as in the great
majority of studies NSO-HETs were also neglected here.
Except for the ltered Ems River sample, suspended particulate
matter (SPM) with a content ranged from 6 to 60 mg L 1 was coextracted. Detailed results concerning concentrations in SPM and
sediment samples will not be presented here but in a forthcoming
publication.
Table 3 shows EQS dened in the WFD for surface waters, insignicance thresholds of NSO-HETs determined for groundwater
(Dieter et al., 2010) and concentrations determined in this study
in the Lower Saxony Campaign. Additionally the compound with
the highest concentration found in each sample is given. In most
cases N-HETs are prevailing, dominated especially by acridine and
quinoline with concentrations of up to 20 ng L 1. As shown, given
maximum allowable concentration EQS values were not exceeded,
whereas the annual average EQS value of the sum of benzo[g,h,i]perylene and indeno[1,2,3-cd]pyrene was exceeded in six samples.
In 2013 the directive 2013/39/EU was adopted by the European Parliament and the Council (European Commission Directive 2013/39/
EU, 2013). It contains revised EQS- and MAC-values for existing
priority substances and 8 newly identied priority substances,
which should be taken into account for the period 20152021. In
sum four substances of the PAHs have to be monitored in water:
anthracene (AA-EQSnew: 100 ng L 1), uoranthene (AA-EQSnew:
6.3 ng L 1), naphthalene (AA-EQSnew: 2000 ng L 1) and benzo[a]pyrene (AA-EQSnew: 0.17 ng L 1). If these new AA-EQS were applied,
in more than 70% of the samples uoranthene and benzo[a]pyrene
concentrations would exceed these threshold values.
3.3.2. River Elbe and River Weser (Second campaign)
In Fig. 4 sum concentrations of EPA-PAHs and NSO-HETs in
the dissolved phase of River Elbe and River Weser are shown. In
the course of the River Elbe no noticeable trends were observed.
The concentrations of EPA-PAHs ranged from 10 to 40 ng L 1 and
for NSO-HETs between 20 and 50 ng L 1 with the exception of
sample E2B. This sample was part of a triplicate measurement
downstream the harbor of Hamburg and had high naphthalene,
methylnaphthalene and dibenzofuran concentrations, which could
not be explained yet.
A collection of worldwide determined concentration levels of
EPA-PAHs in the dissolved aqueous phase can be found in

Fig. 4. Sum concentrations of the EPA-PAHs and the NSO-HETs in the course of
River Elbe and River Weser towards the German Bight, E = River Elbe, W = River
Weser, N = North Sea.

Montuori and Triassi (2012) and for the Elbe River in Heemken
et al. (2000) and ARGE-ELBE (1998). Using the detailed data base
for the Elbe River concentrations in the dissolved aqueous phase
between 9.7 and 19.5 ng L 1 are available for sampling sites downstream Hamburg harbor (ARGE-ELBE, 1998). Concentrations of
EPA-PAHs in the dissolved aqueous phase found in this study correspond to these existing monitoring data in the Elbe River
(excluding sample E2B, see Fig. 4). It has to be noted that concentrations of high molecular EPA-PAHs in the adsorbed state on SPM
are not discussed here. The sum concentrations of NSO-HETs found
in this study in the dissolved aqueous phase were higher than the
EPA-PAHs. Therefore, it is assumed that in inowing smaller rivers
NSO-HETs can be detected with even higher concentrations than
found in this study.
Within the heterocyclic substance group acridine, quinoline and
methylquinolines prevailed (Fig. S2 in supplementary materials).
Acridine concentrations were with about 4 ng L 1 lower than in
the Lower Saxony Campaign and comparable to the maximum concentration in the Spanish River Ebro and its tributaries (LpezSerna et al., 2012).
In contrast to the River Elbe decreasing concentrations towards
the German Bight were found in the River Weser (see Fig. 4). For
example concentrations of the N-HET acridine declined from
7.3 ng L 1 at sampling point W1 in Bremen to 0.7 ng L 1 in the
North Sea and concentrations of the EPA-PAH uoranthene
declined from 2.6 to 0.6 ng L 1. A probable explanation for this
observed trend could be a simple dilution effect. An indicator for
a dilution effect is the increasing salinity towards the German
Bight. Another reason can be partitioning between the aqueous
and solid phase. The suspended matter increases from 12 up to
90 mg L 1 towards the North Sea.
In order to investigate the distribution of NSO-HETs along a riverine course, more data regarding the up-to-now unknown sources
and degradation reactions in surface waters are necessary.
3.3.3. North Sea (Third campaign)
Fluoranthene, 1-indanone, 2(1H)-quinolinone, 9(10H)-arcidinone and 2-methylquinoline were the only substances detected

A.-K. Siemers et al. / Chemosphere 122 (2015) 105114

above LOQ in at least 5 out of 8 North Sea water samples with

concentrations between 0.3 up to 4.4 ng L 1. In 1997 the OSPAR
Commission dened background concentrations for selected substances, such as EPA-PAHs in surface water of the North Sea
(OSPAR Commission, 2000). For uoranthene this background
value is in a range of about 0.10.3 ng L 1 in the central and southern North Sea. In this study up to 1 ng L 1 were found which must,
therefore, be attributed to human activity. Further dened background concentrations of other EPA-PAHs are in a low pg L 1 range.
This could not be achieved with the given LOQs presented in this
study. Lohmann et al. (2009) determined 0.13 ng L 1 uoranthene
off the Norwegian coast. Furthermore most EPA-PAHs investigated
along a North-South Atlantic transect by Nizzetto et al. (2008)
were found in the dissolved phase and were dominated by phenanthrene, uoranthene and pyrene.
It is notable that besides uoranthene as the only EPA-PAH,
three N-HETs were detected in sea-water samples with concentrations in the range of 1 ng L 1. Bester and Theobald (2000) found
xanthone (O-HET) in concentrations between 0.08 and 0.68 ng L 1
in samples from the German Bight. However, these concentrations
are too low to be determined with the LOQ of 2.0 ng L 1 obtained in
this study. 1-Indanone as an assumed degradation product of
indane was found in all samples with concentrations up to 4 ng L 1.
4. Conclusion
In general two different methods were developed to extract a
total of 86 substances containing heterocyclic PAHs, alkylated
PAHs, EPA-PAHs and phenols simultaneously out of water samples.
Although these analytes show a wide range of polarity, LOQs in the
ng L 1 range could be achieved for most of the compounds. Quantication of the majority of substances was realized by GCMS. The
fractionation of the basic N-HETs possesses a suitable clean-up
prior to the more sensitive LCMSMS analysis. The mean recoveries for the robust SPE method were about 80% for the EPA-PAHs,
90% for S-HETs and 80% O-HETs for all three different matrices.
However, recoveries of N-HETs depend on the matrix investigated.
For N-HETs the recovery declined from 88% in ultrapure water to
about 60% in sea water. This study is the rst to examine the occurrence of NSO-HETs in large rivers and the North Sea. Concentrations of NSO-HETs in the range of up to 20 ng L 1 for a single
compound were found in the whole water phase using LLE as
extraction method. In the dissolved aqueous phase using SPE as
extraction method the sum concentrations of NSO-HETs were
higher than the sum concentrations of the EPA-PAHs.
In summary, it can be stated that heterocyclic PAHs occur in
relatively high concentrations in surface waters. In particular, the
N-HETs acridine, quinoline and methylquinolines prevail. More
data regarding the up-to-now unknown sources and degradation
reactions of NSO-HETs are necessary to assess their environmental
concentrations in surface waters.
Acknowledgements
We thank Dieter Steffen for support and we gratefully acknowledge nancial support from the Lower Saxony Water Management,
Coastal Defence and Nature Conservation Agency (NLWKN) for the
rst campaign. Furthermore, many thanks to the HelmholtzZentrum Geesthacht (HZG) for the opportunity to carry out the
second campaign on their research vessel Ludwig Prandtl. In particular thanks to Kerstin Heymann for supporting SPM data and
special thanks to Hendrik Wolschke for organizing the campaign,
creating the map of sampling sites and for providing the seawater
samples. Moreover we acknowledge nancial support from the
Innovation Incubator Lneburg, funded by the European Union

113

allocated by the European Regional Development Fund (ERDF).

The authors would like to thank the anonymous reviewers for their
valuable comments and suggestions to improve the quality of the
paper.
Appendix A. Supplementary material
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.chemosphere.
2014.11.022.
References
Andersson, J.T., Hegazi, A.H., Roberz, B., 2006. Polycyclic aromatic sulfur
heterocycles as information carriers in environmental studies. Anal. Bioanal.
Chem. 386, 891905.
ARGE Elbe, 1998. Verteilung von organischen Kontaminanten zwischen wssriger
und partikulrer Phase in der Elbe und der Deutschen Bucht.
Arbeitsgemeinschaft fr die Reinhaltung der Elbe, Hamburg.
Bester, K., Theobald, N., 2000. Results of non target screening of lipophilic organic
pollutants in the German Bight V: xanthen-9-one. Water Res. 8, 22772282.
Bleeker, E.A.J., van der Geest, H.G., Klamer, H.J.C., de Voogt, P., Wind, E., Kraak,
M.H.S., 1999. Toxic and genotoxic effects of azaarenes: isomers and metabolites.
Polycyclic Aromat. Compd. 13, 191203.
Blotevogel, J., Reineke, A.K., Hollender, J., Held, T., 2008. Identication of NSOheterocyclic priority substances for investigating and monitoring creosotecontaminated sites. Grundwasser 13, 147157.
Brinkmann, M., Maletz, S., Krauss, M., Bluhm, K., Schiwy, S., Kuckelkorn, J., Tiehm, A.,
Brack, W., Hollert, H., 2014. Heterocyclic aromatic hydrocarbons show
estrogenic activity upon metabolization in a recombinant transactivation
assay. Environ. Sci. Technol. 48, 58925901.
Delhomme, O., Millet, M., 2008. Comparison of two analytical methods for the
determination of azaarenes in atmospheric particulate matter. Polycycl.
Aromat. Comp. 28, 518532.
Dieter, H., Frank, D., Herrmann, H., Konietzka, R., Moll, B., Six, E., Stockerl, R., von der
Trenck, K., 2010. Ableitung von Geringfgigkeitsschwellenwerten fr das
Grundwasser. NSO-Heterocyclen. Bund/Lnder Arbeitsgemeinschaft Wasser.
DIN (Deutsches Institut fr Normung), 1994. Method 32645. Chemical Analysis;
Decision Limit; Detection Limit and Determination Limit; Estimation in Case of
Repeatability; Terms, Methods, Evaluation. Beuth Verlag, Berlin.
European Commission Directive 2000/60/EC of European Parliament and of the
Council of 23 October 2000 establishing a framework for Community action in
the eld of water policy. <http://faolex.fao.org/docs/pdf/eur23005.pdf>
(accessed 30.06.14).
European Commission Directive 2008/105/EC of European Parliament and of the
Council of 16 December 2008 on environmental quality standards in the eld of
water policy. <http://faolex.fao.org/docs/pdf/eur84568.pdf> (accessed 30.06.2014)
European Commission Directive 2013/39/EU of European Parliament and of the
Council of 12 August 2013 amending Directives 2000/60/EC and 2008/105/EC as
regards priority substances in the eld of water policy. <http://faolex.fao.org/
docs/pdf/eur127344.pdf> (accessed 30.06.14)
Feldmannova, M., Hilscherova, K., Marsalek, B., Blaha, L., 2006. Effects of Nheterocyclic polyaromatic hydrocarbons on survival, reproduction, and
biochemical parameters in daphnia magna. Environ. Toxicol. 21, 425431.
Gimeno, R.A., Altelaar, A.F.M., Marce, R.M., Borrull, F., 2002. Determination of
polycyclic aromatic hydrocarbons and polycyclic aromatic sulfur heterocycles
by high-performance liquid chromatography with uorescence and
atmospheric pressure chemical ionization mass spectrometry detection in
seawater and sediment samples. J. Chromatogr. A 958, 141148.
Heemken, O.P., Stachel, B., Theobald, N., Wenclawiak, B.W., 2000. Temporal
variability of organic micropollutants in suspended particulate matter of the
River Elbe at Hamburg and the River Mulde at Dessau, Germany. Arch. Environ.
Contam. Toxicol. 38, 1131.
ISO (International Organization for Standardization), 2011. Standard 28540. WaterQuality Determination of 16 polycyclic aromatic hydrocarbons (PAH) in Water
Method using gas chromatography with mass spectromic detection (GCMS).
Beuth Verlag, Berlin.
Jeanneau, L., Faure, P., Jarde, E., 2007. Inuence of natural organic matter on the
solid-phase extraction of organic micropollutants application to the waterextract from highly contaminated river sediment. J. Chromatogr. A 1173, 19.
Johansen, S.S., Hansen, A.B., Mosboek, H., Arvin, E., 1996. Method development for
trace analysis of heteroaromatic compounds in contaminated groundwater. J.
Chromatogr. A 738, 295304.
Kosjek, T., Andersen, H.R., Kompare, B., Ledin, A., Heath, E., 2009. Fate of
carbamazepine during water treatment. Environ. Sci. Technol. 43, 62566261.
Lohmann, R., Gioia, R., Jones, K.C., Nizzetto, L., Temme, C., Xie, Z., Schulz-Bull, D.,
Hand, I., Morgan, E., Jantunen, L., 2009. Organochlorine pesticides and PAHs in
the surface water and atmosphere of the North Atlantic and Arctic Ocean.
Environ. Sci. Technol. 43, 56335639.
Lpez-Serna, R., Petrovic, M., Barcel, D., 2012. Occurrence and distribution of
multi-class pharmaceuticals and their active metabolites and transformation
products in the Ebro River basin (NE Spain). Sci. Total Environ. 440, 280289.

114

A.-K. Siemers et al. / Chemosphere 122 (2015) 105114

Montuori, P., Triassi, M., 2012. Polycyclic aromatic hydrocarbons loads into the
Mediterranean Sea: estimate of Sarno River inputs. Marine Poll. Bull. 64, 512
520.
Mundt, M., Hollender, J., 2005. Simultaneous determination of NSO-heterocycles,
homocycles and their metabolites in groundwater of tar oil contaminated sites
using LC with diode array UV and uorescence detection. J. Chromatogr. A 1065,
211218.
Nizzetto, L., Lohmann, R., Gioia, R., Jahnke, A., Temme, C., Dachs, J., Herckes, P., Di
Guardo, A., Jones, K.C., 2008. PAHs in air and seawater along a North-South
Atlantic transect: trends, processes and possible sources. Environ. Sci. Technol.
42, 15801585.
OSPAR Commission, 2000. Quality Status Report 2000, Region II Greater North
Sea. OSPAR Commission, London, 55 et seqq.
Peddinghaus, S., Brinkmann, M., Bluhm, K., Sagner, A., Hinger, G., Braunbeck, T.,
Eisentrger, A., Tiehm, A., Hollert, H., Keiter, S.H., 2012. Quantitative assessment
of the embryotoxic potential of NSO-heterocyclic compounds using zebrash
(Danio rerio). Reprod. Toxicol. 33, 224232.
Renberg, L., 1981. Gas chromatographic determination of phenolic compounds in
water as their pentauorbenzyl derivates. Chemosphere 10, 767773.
Rostad, C.E., Pereira, W.E., Ratcliff, S.M., 1984. Bonded-phase extraction column
isolation of organic compounds in groundwater at a hazardous waste site. Anal.
Chem. 56, 28562860.
Schfer, R.B., von der Ohe, P.C., Khne, R., Schrmann, G., Liess, M., 2011.
Occurrence and toxicity of 331 organic pollutants in large rivers of north
Germany over a decade (19942004). Environ. Sci. Technol. 45, 61676174.
Schlanges, I., Meyer, D., Palm, W.U., Ruck, W., 2008. Identication, quantication
and distribution of PAC-metabolites, heterocyclic PAC and substituted PAC in

groundwater samples of tar-contaminated sites from Germany. Polycycl.

Aromat. Comp. 28, 320338.
Shinohara, R., Kido, A., Okamoto, Y., Takeshita, R., 1983. Determination of trace
azaarenes in water by gas chromatography and gas chromatography-mass
spectrometry. J. Chromatogr. 256, 8191.
Slobodnik, J., Mrafkova, L., Carere, M., Ferrara, F., Pennelli, B., Schrmann, G., von
der Ohe, P.G., 2012. Identication of river basin specic pollutants and
derivation of environmental quality standards: a case study in the Slovak
Republic. Trends Anal. Chem. 41, 133145.
Steinheimer, T.R., Ondrus, M.G., 1986. Determination of selected azaarenes in water
by bonded-phase extraction and liquid chromatography. Anal. Chem. 58, 1839
1844.
US EPA, 1984. Method 604. Phenols. <https://www.nemi.gov/methods/
method_summary/4789/> (accessed 30.06.14).
US EPA, 1996. Method 3510C. Separatory Funnel LiquidLiquid Extraction. <http://
www.epa.gov/osw/hazard/testmethods/sw846/pdfs/3510c.pdf>
(accessed
30.06.14).
US EPA, 2000. Method 528. Determination of Phenols in Drinking Water by Solid
Phase Extraction and Capillary Column Gas Chromatography/Mass
Spectrometry (GCMS). <https://www.nemi.gov/methods/method_summary/
4677/> (accessed 30.06.14).
US EPA, 2004. Method 3535A. Solid-Phase Extraction (SPE). <https://
www.nemi.gov/methods/method_summary/9187/> (accessed 30.06.14).
US EPA, 2010. Estimation Program Interface Suite for Microsoft Windows (EPISuite). United States Environmental Protection Agency, Washington, DC.