Marine Pollution Bulletin 48 (2004) 341–350

www.elsevier.com/locate/marpolbul

Inhomogeneous distribution of polycyclic aromatic hydrocarbons

in different size and density fractions of contaminated sediment
from Auckland Harbour, New Zealand: an opportunity
for mitigation
Michael J. Ahrens *, Craig V. Depree
National Institute of Water and Atmospheric Research (NIWA), P.O. Box 11-115, Gate 10, Silverdale Road, Hamilton, New Zealand

Abstract

Polycyclic aromatic hydrocarbons (PAHs) in sediment from Auckland Harbour (New Zealand) are not distributed evenly
throughout bulk sediment, but highly concentrated in coarser, low-density fractions. Concentrations of 24 PAHs, measured in
sediment that was separated into six size fractions that were furthermore separated into two density sub-fractions by flotation in
sodium-polytungstate solution (q ¼ 2:15 g cm3 ), varied between 4–103 lg g1 dw among grain size fractions and 2–998 lg g1 dw
for density sub-fractions. Highest PAH concentrations were measured in the low density, 125–250 lm fraction. All sediment
fractions had a similar relative PAH composition, dominated by >3-ring PAHs, suggesting a common pyrogenic origin. Low density
material had 10–200 times higher PAH concentrations and 10–100 times higher organic carbon (OC) content, yet differences in OC
content only partially accounted for variations in PAH concentration. Low density particles contributed more than 75% of the
RPAH, while comprising only 3% of bulk sediment dry weight. This may have significant utility for contaminant mitigation efforts in
Auckland Harbour.

2003 Elsevier Ltd. All rights reserved.

Keywords: PAHs; Sediment; Grain size; Density; Organic carbon; Auckland harbour

1. Introduction straightforward and effective for screening out sedi-

Environmental regulators and managers are repeat-
edly faced by the questions: how contaminated is a
sediment, does it exceed the sediment quality guideline
(SQG), and can we clean it up? Concentrations of
polycyclic aromatic hydrocarbons (PAHs) in urban
sediments vary by several orders of magnitude (Mc-
Cready et al., 2000; Wang et al., 2001; Fang et al., 2003)
and often exceed SQGs designed to protect aquatic
biota and, ultimately, ourselves. Identifying which sed-
iments are of concern and which ones are not is, there-
fore, of crucial importance for directing cost-effective
mitigation efforts. However, monitoring programmes
that measure PAH contamination in aquatic sedi-
ments routinely report bulk PAH concentrations refer-
enced to sediment dry weight. While methodologically
*
Corresponding author. Tel.: +64-7-856-1730; fax: +64-7-856-0151.
E-mail address: m.ahrens@niwa.co.nz (M.J. Ahrens).
0025-326X/$ - see front matter
2003 Elsevier Ltd. All rights reserved.
doi:10.1016/j.marpolbul.2003.08.013
ments whose PAH concentrations fall well below SQG
‘‘trigger’’ values, this approach has a number of con-
ceptual shortcomings when dealing with contaminated
samples. The first is the tacit assumption that a mean
PAH concentration adequately represents the PAH
contamination of a sediment sample. However, sedi-
ment is typically not a uniform medium, but a hetero-
geneous mixture of particles of different origin, grain
size and physicochemical properties. More often than
not, PAHs are concentrated in specific particle fractions
while largely absent in others. Thus, even if a PAH
measurement can be reproduced in several replicate
measurements, this procedural precision is no guarantee
that a bulk concentration value is representative of the
PAH contamination of the sediment particle ‘‘popula-
tion’’. For example, in a sediment with a bimodal dis-
tribution of PAHs among different particle fractions, a
bulk PAH concentration would describe an essentially
nonexistent, hypothetical particle with intermediate
contamination, whereas few particles would actually
342 M.J. Ahrens, C.V. Depree / Marine Pollution Bulletin 48 (2004) 341–350
have this ‘‘average’’ property. Inaccurately inferring
individual particle properties from bulk sediment
properties can have severe implications for models pre-
dicting delivery of contaminated sediment or ingestion
by benthic biota. Secondly, bulk, dry-weight referenced
PAH concentrations allow only very limited general-
izations about PAH ‘‘mobility’’, such as bioavailability
to microbes or sediment-dwelling macrofauna. PAHs do
not bind strongly to inorganic sediment grains, but
preferentially adsorb to hydrophobic organic carbon
(OC) domains (Karickhoff et al., 1979). This is the
common explanation of why normalization to bulk OC
content tends to reduce differences in PAH concentra-
tions between samples, and improves correlations with
organism tissue burdens (Di Toro et al., 1991). Never-
theless, even when normalizing to bulk sediment OC,
PAH concentrations in pore waters or tissues of benthic
biota often deviate by over two orders of magnitude
from predictions (Hickey et al., 1995; Meador et al.,
1995; Burgess et al., 2003). Furthermore, it is commonly
observed that only a fraction of the measurable PAHs
equilibrate rapidly with surrounding pore waters or or-
ganism tissues, while the remaining PAH pool takes
much longer or does not equilibrate at all (Huang et al.,
1997; Cornelissen et al., 1998). The observed incomplete
or time-dependent PAH desorption appears to be
caused by differences in the sorption strength of PAH
molecules to organic matter, and once again invokes
sediment heterogeneity: whereas bulk OC-normalization
implicitly assumes that PAHs partition uniformly
throughout the entire sediment OC fraction, numerous
recent studies have shown that PAHs preferentially as-
sociate with specific OC types or
domains’ within the
OC pool (Gustafsson et al., 1997; Cuypers et al., 2000;
Rockne et al., 2002; Ahrens and Hickey, 2003). For
example coal sediment components have been found to
have higher PAH concentrations than organic coatings
on inorganic sediment particles (Ghosh et al., 2000), and
low density, OC-rich fractions can contain up to 80% of
a bulk sediment’s PAH loading (Rockne et al., 2002).
Furthermore, sediments with large amounts of soot or
other forms of black carbon (BC) show greatly reduced
PAH bioavailability (McGroddy and Farrington, 1995;
McGroddy et al., 1996; Gustafsson and Gschwend,
1997).
The question thus arises, whether one can identify
a sediment sub-fraction that more accurately repre-
sents the PAH contamination and bioavailability of a
sediment sample than its respective bulk sediment con-
centration? Unfortunately, generalizations about bio-
availability and distribution of PAHs that may apply
to one sediment sample often do not apply to another,
even in nearby locations. For example, the strongly
engrained conception that PAHs tend to be concen-
trated in the fine-grained (<63 lm) sediment fraction,
supposedly reflecting higher OC content and greater
surface area, is only true in a subset of cases. For ex-
ample, Karickhoff et al. (1979) and Maruya et al. (1996)
found PAH concentrations to be positively correlated
with a greater proportion of fines (<63 lm). However, in
more than a few exceptions, the highest PAH concen-
trations and inventories have been found on coarser
particle fractions (Prahl and Carpenter, 1983; Simpson
et al., 1998; Wang et al., 2001; Rockne et al., 2002).
Evans et al. (1990) even observed a bimodal distribution
of PAHs and OC in grain size fractions of river sedi-
ments, and Rockne et al. (2002) found PAHs to be
highly enriched in low-density, coarse-grained fractions
that comprised only a small proportion of the bulk
sediment mass. Thus, if one generalization appears to be
valid, it is that PAH distributions in sediments are no
less heterogeneous than distributions of grain size, OC
content and OC composition. Accurate modeling of
contamination inputs, therefore, necessitates the ap-
propriate characterization of contaminant distribution
among different particle fractions. Subsequent mitiga-
tion efforts will be most effective if they focus on the
most contaminated and most bioavailable sediment
fractions. An ideal mitigation scenario would have
contaminants highly concentrated in a small sub-frac-
tion of the sediment that is easily separable from the
remaining bulk sediment.
A further uncertainty for transport modeling of
contaminated sediments is whether the observed con-
taminant distributions in sediment fractions reflect those
of the source material, or, instead, represent the out-
come of post-depositional re-distribution processes. For
example, did a highly PAH-contaminated sediment
particle acquire its PAH ‘‘signature’’ on shore or after it
entered the water body? In the former case, particle in-
terception prior to deposition would make sense for
mitigation, while in the latter case it would have little
effect. If PAH concentrations vary greatly among dif-
ferent sediment fractions, what can be said about their
relative composition? Do sediment fractions that have
highly variable PAH concentrations also show markedly
different compositional signatures? Since different PAH
compounds partition into organic matter depending on
their relative fugacity, it is likely that the relative com-
position of the PAHs analyzed will be uniform among
various particle sizes and classes. A finding of uniform
PAH profiles among different size fractions can, there-
fore, be taken as evidence for a common source or, al-
ternatively, for the presence of dynamic exchange and
equilibration processes within the sediment matrix.
The purpose of this paper is to examine PAH distri-
butions within fractions of a contaminated sediment
from Auckland Harbour, New Zealand. We were in-
terested in whether PAHs would be concentrated in
specific sub-fractions that could subsequently be tar-
geted for mitigation action. In addition to examining the
distribution of PAHs and OC among six grain size
 M.J. Ahrens, C.V. Depree / Marine Pollution Bulletin 48 (2004) 341–350
fractions, we furthermore analyzed their distribution in
two sub-fractions of different density: an easily floatable,
OC-rich fraction, and a heavy, primarily inorganic
fraction. We present evidence that PAH concentrations
vary by several orders of magnitude even among frac-
tions of the same sediment sample, and that it is not
necessarily the smallest particle size or most abundant
sediment fractions that contain the majority of PAHs.
2. Materials and methods
2.1. Study site and sample collection
Sediment used in this study was collected from Mo-
tions Creek (MC), a 2 km long tidal inlet located in
Waitemata Harbour within the city limits of Auckland,
New Zealand, in November 2001. MC sediment had a
soft, sticky consistency, with a predominance of silt-
sized particles. Mud crab burrows abounded at the
sampling location. The inlet is fringed by mangroves
(Avicennia marina), which give way to coarser-grained
sand flats as the creek empties out into Waitemata
Harbour, which comprises the eastern portion of
Auckland Harbour. MC receives storm water from a
nearby motorway, and historically also received dis-
charge from a slaughterhouse, a tallow works and two
tanneries. Previous work by us has established that
sediments contain elevated levels of PAHs and Cu and
Zn (unpubl. obs.). Sediment was collected at low tide
using a PVC core tube (£ 13 cm). Cores were sectioned
in situ into a surface layer (0–3 cm) and a deep layer (10–
13 cm). To assure greater homogeneity, sediment from
eight cores was combined, rendering a total sediment
wet mass of 3.5–4.5 kg, respectively. We report here
results for the surface layer only.
2.2. Sediment size separation
Collected sediment was kept refrigerated (4 C) for 1d
before homogenization and sieving. Samples were wet-
sieved (sea water, 34 ppt) through a series of sieves, to
render 6 size fractions: >1000, 500–1000, 250–500, 125–
250, 63–125 and <63 lm. Particles in the <63 lm frac-
tion were concentrated by centrifugation (4200 · g, 15
min). Sub-samples of all size fractions were lyophilized,
and total sediment dry weight was back-calculated from
a wet:dry weight ratio, determined for a sub-sample of
bulk sediment.
2.3. Sediment density separation
To assess PAH distribution between the sediment-
bound OC pool and an easily floatable OC fraction,
sediment fractions were furthermore separated into two
density sub-fractions. For this purpose, sub-samples (10–
343
30 g) of wet-sieved grain size fractions were centrifuged
(825 · g, 10 min) to remove porewater. Samples were
subsequently resuspended in approximately 40 ml of
sodium-polytungstate solution (3Na2 Æ (WO4 Æ 9WO3 ) Æ
H2 O), with a density of 2.15 g cm3 (TC-Tungsten
Compounds, Germany), shaken vigorously, and re-cen-
trifuged (470 · g, 10 min). The floating particulate sub-
fraction obtained thus was operationally termed ‘‘light’’
(density < 2.15 g cm3 ) whereas the settled particulate
fraction was termed ‘‘heavy’’ (density > 2.15 g cm3 ).
The light fractions were isolated by filtration (Whatman
No. 4), rinsed with deionized water and dried at 40 C.
Dry-weight percentages were calculated by converting
the wet weight of starting material to dry weight, using
the previously established wet:dry weight ratio.
2.4. POC analysis
Sediment particulate organic carbon (OC) of each
size fraction and density sub-fraction was determined by
combustion on a ThermoQuest NC 2500 elemental an-
alyzer. Sediment samples were homogenized, acidified
with H2 SO4 to remove carbonates, and dried at 60 C
before analysis.
2.5. Analytical procedures
Organic solvents were distilled at least once prior to
use. The deuterated PAH surrogates, naphthalene-d8 ,
acenaphthene-d10 , phenanthrene-d10 , fluoranthene-d10 ,
pyrene-d10 , benz(a)anthracene-d12 and perylene-d12 were
purchased individually from Supelco (Bellefonte, PA,
USA). The calibration mixture consisting of 24 PAHs
(SRM2260) was purchased from NIST, which com-
prised the 16 US EPA priority PAHs and, furthermore,
1-methylnaphthalene, 2-methylnaphthalene, biphenyl, 2,6-
dimethylnaphthalene, 2,3,5-trimethylnaphthalene, 1-meth-
ylphenanthrene, benzo(e)pyrene, and perylene. Silica gel
(Silica gel 60, 0.063–0.2 mm, Merck, Darmstadt, Ger-
many), alumina (F-20, 80-200 mesh, Sigma-Aldrich,
Auckland, New Zealand) and Na2 SO4 used for clean-up
columns was heated for at least 16 h at 170, 130 and 600
C respectively and stored in a dessicator prior to use.
Cotton wool used as packing support for column chro-
matography was cleaned by Soxhlet extraction with di-
chloromethane (DCM) for 18 h. Copper granules were
activated by concentrated HCl treatment (5 min), fol-
lowed by rinsing with 2· methanol and 3· DCM (and
stored in the latter) immediately prior to use. All glass-
ware was thoroughly washed, rinsed with deionized
water and baked at 450 C overnight, prior to use.
2.6. PAH extraction and fractionation
PAHs were extracted with DCM using an accelerated
solvent extractor (Dionex ASE-200). One gram of
344 M.J. Ahrens, C.V. Depree / Marine Pollution Bulletin 48 (2004) 341–350

freeze-dried, homogenized sediment of each grain size Table 1

fraction, or 0.01–1.0 g of sediment from the density sub- Sediment characteristics of Motions Creek sediment (0–3 cm)
fractions, were weighed directly into 22 ml stainless steel Grain size Mass con- Corg % RPAH RPAHOC
extraction cells and spiked with an appropriate amount (lm) tribution (lg/g dw) (lg/g OC)
to total dry
of a surrogate solution containing naphthalene-d8 , ace- wt
naphthene-d10 , phenanthrene-d10 , fluoranthene-d10 , py-
>1000 0.3% 36.8 5.2 14
rene-d10 , benz(a)anthracene-d12 and perylene-d12 . 500–1000 0.4% 32.3 67.4 209
Extracts were transferred to a pre-conditioned (10 ml 250–500 1.6% 11.1 103.1 929
DCM) glass column (160 mm · 12 mm), containing a 125–250 25.3% 1.4 24.0 1725
plug of cotton wool and dry-packed with alumina (1 g), 63–125 16.5% 1.0 14.8 1450
<63 56.0% 2.4 4.0 166
silica gel (4 g) and anhydrous Na2 SO4 (1 g). After elu-
tion with DCM (30 ml), the sample volume was reduced Compositea 2.3 12.6 557
via rotary evaporation (B€ uchi Labortechnik AG, Flawil, bulk 2.0 16.6 812
a
Switzerland), exchanged into heptane (ca. 300 ll), after Composite reflects the mass-weighted average of all grain size
which activated Cu granules were added to remove fractions.
sulfur. Prior to transfer to GC vials (350 ll inserts),
sample extracts received 500 ng of internal standard (p-
terphenyl-d14 ; 10 lg ml1 ). pyrene-d10 (87 ng ll1 in methanol), giving corre-
PAHs were analyzed on a Hewlett Packard 5890 Se- sponding aqueous pyrene-d10 concentrations of 44, 87
ries II gas chromatograph (GC), coupled to a Hewlett and 130 lg l1 , respectively. All concentrations were
Packard 5972 mass spectrometer (MS), operated in se- below pyrene’s aqueous solubility of 135 lg l1 . One
lected ion monitoring (SIM) mode. The GC was gram of sediment sample was then added to each vial
equipped with a split/splitless injector, 7673 auto-injec- (the 125–250 lm sediment was first ground by mor-
tor and 30 m · 0.25 mm (0.25 m film thickness) ZB-5 tar and pestle), after which the centrifuge tubes were
column (Phenomenex, Auckland, New Zealand). GC capped and agitated on an orbital shaker for 8 d at
operating conditions were as follows: inlet and transfer approximately 20 C. All treatments were carried out
line temperature both 280 C; 1 ll splitless injection with in duplicate. Aqueous pyrene-d10 concentrations were
1 min purge delay; 0.7 ml min1 flow rate of He carrier determined as follows: samples were centrifuged at
gas; initial column temperature of 60 C, increased to 2500 · g, after which a 20 ml aliquot was removed,
130 C at 20 C min1 and then at 4 C min1 to a final spiked with surrogate (fluoranthene-d10 ) and extracted
temperature of 290 C, with a final holding time of 10 with hexane (2 · 5 ml). The extract was dried (MgSO4 ),
min. Identification of the 24 target PAHs was based on a reduced to 1 ml by rotary evaporation, and an internal
combination of retention time and comparison of rela- standard (p-terphenyl-d14 ) was added prior to trans-
tive ion abundances with authentic standards. Quanti- ferring to a GC vial. Particulate concentrations were
fication was performed in SIM mode, with relative determined as follows: after separation from the aque-
response factors (relative to the internal standard) of the ous-phase, particulate material was lyophilized, spiked
24 PAHs determined from authentic standards during with surrogate, and subjected to accelerated solvent
instrument calibration, which consisted of at least a six- extraction with DCM. Sediment PAH extracts were then
point calibration curve. PAH concentrations were cor- worked-up as following the standard procedure de-
rected for surrogate recovery. Surrogate recoveries scribed earlier. The concentrations of pyrene-d10 in
ranged between 77–113%. aqueous and particulate phases were quantified by
GCMS (SIM mode) and corrected for fluoranthene-d10
2.7. PAH sorption measurements recovery. Concentrations of pyrene-d10 in particulate
fractions were plotted versus corresponding concentra-
To compare the binding affinity of different grain tions in the aqueous fraction, and values of log KOC were
size fractions, a small-scale experiment was conducted derived by linear regression, after normalizing for OC
to measure the partitioning of pyrene between aque- content, as given in Table 1.
ous phase and three particulate fractions: bulk MC
sediment, and the 125–250 lm and <63 lm grain size
fractions. Because of the high levels of pyrene contam- 3. Results and discussion
ination, the sorption studies were conducted using
deuterated pyrene (pyrene-d10 ) as the probe and fluo- 3.1. Sediment characteristics
ranthene-d10 as a surrogate. For each of the sediment
samples, three glass centrifuge tubes (45 ml, CorexII, Grain size characteristics of MC sediment are sum-
with Teflon-lined screw-cap) containing 40 ml of de- marized in Table 1. Grain size distribution, as compared
ionized H2 O were spiked with 1740, 3480 or 5220 ng of by dry weight, was bimodal, with maxima in the sand
 M.J. Ahrens, C.V. Depree / Marine Pollution Bulletin 48 (2004) 341–350 345

(125–250 lm; 25%) and silt (<63 lm; 56%) fractions.
Particulate organic carbon (OC) content, as measured
by elemental analysis, ranged between 1–37% and
showed a positive correlation with grain size, with
highest OC content measured in the coarser fractions.
Wood and plant debris were prominent components in
the coarse grain size fraction, while only isolated inor-
ganic particles such as bivalve shell fragments and a
small number of pebbles were found. The mass-weighted
contribution of each grain size fraction to total (com-
posite) sediment OC content ranged between 10% and
20%, except for the finest (<63 lm) grain size fraction,
which contributed 35%. Results for surface (0–3 cm) and
deep (10–13 cm) sediment layers were very similar for
grain size distribution, OC content, and concentration
range and composition of PAHs, and, consequently, are
presented for the surface sediment layer only.
3.2. Distribution of PAHs in different size fractions
The PAH composition of MC sediment was domi-
nated by pyrene and fluoranthene and elevated con-
centrations of >4-ringed PAHs (Fig. 1). Concentrations
of two and three-ringed PAHs were typically 10 times
lower, except for phenanthrene, which was the only
abundant three-ringed PAH. The relative PAH com-
position was remarkably similar between the six grain
size fractions analyzed: fluoranthene and pyrene gener-
ally contributed between 10% and 20% of the sum-total
of the 24 PAHs analyzed (RPAH ¼ sum of 16 EPA + 8
other PAHs), while phenanthrene and 4–5 ring PAHs
each contributed on the order of 5–10%, respectively.
While the relative PAH composition (termed ‘‘PAH
signature’’ for the remainder of this study) was largely
identical between different grain size fractions, absolute
concentrations varied by more than 10-fold for indi-
vidual compounds and RPAH. Highest PAH concen-
trations occurred in the intermediate grain size fractions
(125–250, 250–500, 500–1000 lm, Fig. 1), while the
coarsest and finest fractions had the lowest PAH con-
centrations. Summed PAH concentrations for the six
particle size classes analyzed, summarized in Table 1,
showed a range for RPAH of 4–103 lg g1 dry weight
(dw). In comparison, bulk sediment RPAH concentra-
tion was 16.6 lg g1 dw, a value which exceeded the
current New Zealand interim sediment quality guideline
of 4 lg g1 dw. Normalization of weight-referenced
PAH concentrations to OC content further amplified
differences in PAH concentration between grain size
fractions (range: 14–1725 lg g1 OC) and shifted the
PAH concentration maximum from the 250–500 lm
fraction to the 125–250 lm fraction (Fig. 2).
To calculate the contribution of each size fraction to
the total (composite) sediment PAH concentration,
RPAH concentrations for each size fraction were multi-
plied by their proportion of sediment dry mass. This
procedure revealed the 125–250 lm grain size fraction to
accommodate the largest proportion of the RPAH con-
centration in MC sediment (48%, Fig. 3), while it com-
prised only 25% of the sediment mass. In contrast, the
abundant <63 lm fraction, which accounted for over 50%
of the sediment mass, contributed merely 17% of RPAH.
3.3. Sediment characteristics and distribution of PAHs in
different density fractions
Grain-size fractions of MC sediment were further di-
vided into two density fractions by flotation in solution

100000
>1000
500-1000
250-500
125-250
10000
63-125
<63
PAH (ng/g dw)

bulk
1000

100

10

1
H
AC PH
-N H

2, iph H

AP

FL E
Y

RY
2M PH

M n

M E

A
TH

P
1M T

A
E

R
F
F

P
1M AP

B P

P
FL

H
TH
A
6D e

3, AC

AN

DB

hi
Bk
PY

Ba

Bb
PH

PE
Be

Ba
A

In
-N

-P
-N

CH
NA

Bg
-N

5T
2,

Fig. 1. Concentrations of 24 PAHs in Motions Creek sediment (abbreviations in parentheses): naphthalene (NAPH); 2-methylnaphthalene (2M-
NAPH); 1-methylnaphthalene (1M-NAPH); biphenyl (Biphen); 2,6-dimethylnaphthalene (2,6DM-NAPH); acenaphthylene (ACTHY); acenaphth-
ene (ACE); 2,3,5-trimethylnaphthalene (235TM-NAPH); fluorene (FL); phenanthrene (PHE); anthracene (ANT); 1-methylphenanthrene (1M-PHE);
fluoranthene (FLTH); pyrene (PYR); benz[a]anthracene (BaA); chrysene (CHRY); benzo[b]fluoranthene (BbF); benzo[k]fluoranthene (BkF);
benzo[e]pyrene (BeP); benzo[a]pyrene (BaP); perylene (PER); indeno[1,2,3-c,d]pyrene (InP); dibenz[a; h]anthracene (DBA); benzo[g; h; i]perylene
(BghiP).
346 M.J. Ahrens, C.V. Depree / Marine Pollution Bulletin 48 (2004) 341–350

10000

contribution to total

0.04%
0.04%
0.69%
Heavy
PAH ug/g dw

22.3%
23.9%
7.9%
5.1%
8.5%
PAH ug/g OC

Mass-weighted
PAH µg/g dw or µg/g Corg

1000

0.06%

11.2%
42.1%
14.2%

77.7%
76.3%
1.3%

8.9%
Light
PAH
100

RPAH (lg per g OC)

Heavy
430
220
951
791

212
241
94
1425
10

Corg normalized
1

Light

1296
2786
2040

1336
1741
171

634
10
3
0

lk
25
0

0
00

<6
00

50

25

bu
-1
10
>1

0-

5-

63
0-

25

12
50

grain size (µm)

RPAH (lg per g dw)

Heavy
16.0
6.9
8.7
4.9
4.8
2.4
3.5
3.8
Fig. 2. Totals of 24 PAHs (RPAH) for bulk and size-fractionated
Motions Creek sediment. Light bars: RPAH normalized to dry weight,
dark bars: RPAH normalized to organic carbon content.

Light

63.5
465
998
704
138
416
373
3.8
80%
dry weight

contribution to total

0.06%
Heavy
sum PAH

0.1%
0.4%
6.1%
2.2%
55.5%
64.4%
69.1%
contribution to total mass (%)

POC

Mass-weighted
60% Sediment characteristics of Motions Creek surface sediment, separated by grain size and density (2.15 g/cm3 )

sediment POC

35.6%
30.9%
3.7%
4.5%
5.3%
9.3%
4.3%
8.6%
Light
40%
Heavy

20%
0.91%
0.62%
0.34%
3.7%
3.1%

2.5%
1.7%
1.6%
POC (%)

0%

Composite reflects the mass-weighted average of all grain size fractions.

37.4%
37.2%
35.9%
35.8%
34.5%
21.8%
31.1%
21.4%
>1000 500-1000 250-500 125-250 63-125 <63
Light

grain size fraction (µm)

Fig. 3. Relative contribution of individual grain size fractions to

contribution to total

composite sediment dry mass, total PAH concentration, and particu-

Heavy

24.6%
16.1%
55.0%
97.1%
96.8%
0.0%
0.1%
1.2%

late organic carbon content (POC).

Mass-weighted

sediment dw

of sodium-polytungstate, to investigate the PAH distri-

0.25%
0.30%
0.37%
0.65%
0.31%
0.99%
Light

2.9%
3.2%

bution between an easily floatable versus a sediment-

bound OC pool. The two density sub-fractions thus
separated were termed light (L, q < 2:15 g cm3 ) and
Heavy
14.4%
23.3%
76.9%
97.4%
98.1%
98.2%

96.8%
Mass contribution

heavy (H, q > 2:15 g cm3 ). Table 2 summarizes the

relative contribution of L and H sub-fraction to size
to fraction dw

fraction mass and to total (composite) sediment mass.

As noted before, surface sediment and deep sediment
85.6%
76.7%
23.1%
2.6%
1.9%
1.8%

3.2%
Light

samples showed very similar quantitative and qualitative

trends in mass distribution, OC content and PAH sig-
nature. For assuring mass balance, bulk sediment that
fraction (lm)

Compositea
Grain size

was only density-fractionated but not grain size frac-

500–1000
250–500
125–250
63–125

tionated was analyzed as well. The contribution of the L

>1000
Table 2

bulk
<63

sub-fraction to the respective size fraction mass ranged

a

between 2% and 86% and was highly correlated with

 M.J. Ahrens, C.V. Depree / Marine Pollution Bulletin 48 (2004) 341–350
grain size: In coarser grain size fractions, the L sub-
fraction contributed a greater proportion of the fraction
mass. Summing the mass-weighted contributions for all
L-sub-fractions, revealed these to comprise 2.9% of the
composite sediment mass. This mass-balance agreed
closely with results for bulk sediment that had only been
density-separated but not sieved: for this, the L fraction
contributed 3.2% of the sediment mass.
Organic carbon content differed markedly between L
and H sub-fractions: While OC content of L sub-frac-
tions ranged between 22% and 37%, OC content of H
sub-fractions was 10–100 times lower, ranging between
0.3% and 3.7%. For both L and H sub-fractions, OC
content was highest in the coarsest grain fractions and
declined steadily with diminishing grain size, suggesting
a loss of OC through attrition processes. The only ex-
ception to this trend was the H (<63 lm) sub-fraction,
which had an OC similar to coarser particle size frac-
tions (2.5%). The ratio of light OC (L-OC) to heavy OC
(H-OC) increased from 10 to 100 with diminishing grain
size. This suggests that organic carbon coatings on
heavier sediment material contributed increasingly less
to sediment OC content in smaller grain size fractions.
However, it is impossible to state whether this obser-
vation reflected a lower OC coverage per surface area,
since we did not determine specific surface area, and it is
quite possible that particle geometries varied in different
grain size fractions.
PAHs were 10–200 times more concentrated in L sub-
fractions compared to H sub-fractions, with the excep-
tion of the coarse (>1000 lm) fraction, which showed
the opposite pattern (Table 2). While PAH concentra-
tions ranged from approximately 2–16 lg g1 dw in the
H-fraction, they ranged from 4–998 lg g1 dw in the L
sub-fraction. Highest PAH concentrations were found
in the 125–250 lm L-sub-fraction, which had nearly 200
times higher PAH concentrations than the respective
125–250 lm H-sub-fraction, and had nearly 30 times
higher PAH concentrations than the non-density sepa-
rated 125–250 lm grain-size fraction (Table 1). Mass-
weighted proportions of L and H sub-fractions to
RPAH of composite sediment, showed over 40% of
RPAH to be associated with the light 125–250 lm sub-
fraction alone (Table 2 and Fig. 4), while the remaining
L sub-fractions contributed another 35%. Summed to-
gether, L-sub-fractions thus contributed 78% of RPAH
of the composite sediment, while only comprising ap-
proximately 3% of the sediment dry weight. A nearly
identical percentage was found for bulk sediment (76%),
which had only been density separated but not sieved.
These results were very similar to the results of Rockne
et al. (2002), who found 50–80% of PAHs to be asso-
ciated with low density (<1.7 g cm3 ) material in sedi-
ment samples from the New York/New Jersey Harbor.
Similarly, this light material contributed only 3–15% of
sediment mass. It should be noted that while relative
347
100%
Light
Heavy
% of composite ΣPAH
80%
60%
40%
20%
0%
>1000 500-1000 250-500 125-250 63-125 <63 bulk
Fig. 4. Relative contribution of light (L, q < 2:15 g cm3 ) and heavy
(H, q > 2:15 g cm3 ) density sub-fractions to total sediment PAH
concentration. Un-sieved, density-separated bulk sediment is shown
for comparison.
proportions and total RPAH in grain size and density
separated sediment were practically identical to density-
separated bulk MC sediment (11.1 vs. 11.2 ng g1 , re-
spectively), sediments having undergone the density
separation step contained only 67% of the RPAH of
non-density separated bulk sediment (16.6 ng g1 , Table
1). However, grain-size fractionated sediment also con-
tained only about 76% RPAH of bulk sediment (12.6
ng g1 , Table 1). This suggests substantial desorption of
PAHs during either the density-flotation or sieving
procedure.
Since L-sub-fractions contained 10–100 times higher
OC than the H sub-fractions, normalization of RPAH
concentrations to OC content diminished differences in
PAH concentrations between density-separated sub-
fractions considerably. The L-sub-fraction in coarser
grain size fractions (>500 lm) was found to be relatively
depleted in RPAHs compared to the corresponding H
sub-fraction (Fig. 5). In contrast, RPAH concentra-
tions in L sub-fractions for all finer grain sizes (<500
lm) were up to seven times more enriched relative to H-
sub-fractions, mirroring results for bulk (non-size–frac-
tionated) sediment, for which OC-normalized RPAH
concentrations were also seven times more concentrated
in the L sub-fraction. These findings demonstrate that
the higher OC content in the L sub-fractions was not the
exclusive explanation for the L-sub-fraction’s higher
contribution to total PAH. The results rather suggest
that, in addition to being more enriched in OC, the
majority of the L sub-fraction’s organic carbon also had
a higher binding affinity for PAHs than H-OC.
L and H sub-fractions had a similar signature of in-
dividual PAHs, with high relative abundances of fluo-
ranthene, pyrene, phenanthrene and >4 ring PAHs, in
keeping with the compositional profile of non-density
separated bulk sediment. Individual PAH concentrations
348 M.J. Ahrens, C.V. Depree / Marine Pollution Bulletin 48 (2004) 341–350

3000 3.4. Sources and sorption of PAHs in Motions Creek

Light
sediment
2500 Heavy
The general similarity of PAH compositional profiles
PAH µg/g orgC

2000 among grain size and density fractions strongly suggests

1500
1000
500
0
>1000 500-1000
Fig. 5. PAH concentrations (OC-normalized) for density and size
separated sub-fractions. Un-sieved, density separated bulk sediment is
shown for comparison.
in the L and H sub-fractions of each grain size were
generally closely correlated, and concentrations of in-
dividual PAH compounds (OC normalized) typically
agreed within a factor of 10 for corresponding L and H
sub-fractions (Fig. 6). One exception was the coarsest
grain size fraction (>1000 lm), for which individual
PAH concentrations in the L-sub-fraction were, on av-
erage, 30 times lower than in the H sub-fraction. The L-
OC of finer grain size fractions (<500 lm) was typically
more enriched in PAHs relative to H-OC, especially for
3–5 ring PAHs. The same PAH enrichment pattern was
also seen for the bulk L-OC (Fig. 6).
a common PAH source. The predominance of >3 ring
PAHs, especially fluoranthene and pyrene, and the rela-
tively low abundance of alkylated naphthalenes and
phenanthrenes furthermore suggest a pyrogenic origin.
Very similar PAH signatures have been reported for
Sydney Harbour (McCready et al., 2000) and Boston
Harbor (Wang et al., 2001), which implicates a similar
250-500 125-250 63-125 <63 bulk
source for sediment-associated PAHs in urbanized es-
tuaries. Urban runoff is generally considered the major
grain size (µm)
source of PAHs entering urbanized estuaries. For ex-
ample, Hoffman et al. (1984) calculated that 71% of the
high molecular weight PAHs entering Narragansett Bay
(USA) were attributable to urban runoff. However,
source apportionment studies on MC sediments indicate
that while there are petroleum-type inputs consistent
with road dust and atmospheric fallout (i.e. urban run-
off), the main PAH contaminant phase appears to be
historic coal-tar related materials (Depree and Ahrens,
2003).
Despite the similarity in PAH composition, absolute
PAH concentrations differed markedly among grain size
fractions and among density sub-fractions. This hetero-
geneity persisted even after normalizing PAH concen-
trations to OC content, suggesting that PAH binding
affinity of sediment OC is not uniform, and that nor-
malization to bulk OC may mask substantial differences
in PAH concentration on different particles. While the

12

10
L/H PAH ratio (OC basis)

8 >1000
500-1000
250-500
6 125-250
63-125
<63
bulk L/H
4

0
H PH PH en PH HY E PH FL E T E R Y F F T P
AP A T AC A H AN H LTH PY BaA HR Bb Bk BeP BaP PE In BA iP
A ph A P -P F D gh
N -N -N Bi -N AC -N C B
2M 1M M M 1M
,6 D , 5T
2 3
2,

Fig. 6. Ratio of individual PAHs (OC-normalized) in low density (L) to high density (H) sediment fractions, showing a relative enrichment of PAHs
in finer sediment OC fractions. For abbreviations of individual PAHs refer to Fig. 1.
 M.J. Ahrens, C.V. Depree / Marine Pollution Bulletin 48 (2004) 341–350
observed concentration differences could also be ex-
plained by insufficient equilibration time between the
various size and density fractions, it seems more likely
that these are due to variable sediment OC composition
and PAH binding affinity. Visual inspection of size and
density separated sediment fractions confirmed the
presence of a multitude of different organic material,
including wood, bark, plastic and a number of different
forms of black carbon (BC).
The presence of BC facies led us to hypothesize that a
large amount of PAHs may be associated with con-
densed soot phases, on account that these materials are
found abundantly in urbanized sediments and typically
exhibit 200 times higher partition coefficients for PAHs
compared to generic, humic-type sediment OC (McGr-
oddy and Farrington, 1995; Gustafsson et al., 1997).
For this reason, we decided to measure the strength of
PAH sorption, by determining the sediment partitioning
coefficient (log KOC ) of spiked, deuterated pyrene for
bulk MC sediment and the two fractions with the
highest and lowest OC-normalized PAH concentrations
(disregarding the >1000 lm fraction). As shown in
Table 1, bulk sediment contained 812 lg g OC1 of PAHs,
while the 125–250 lm and <63 lm fractions contained
1725 and 166 lg g OC1 , respectively. On this basis, we
hypothesized that the 125–250 lm fraction might con-
tain more BC and the <63 lm fraction less BC relative
to bulk sediment, and that these differences would be
reflected in differences in the pyrene partitioning coeffi-
cient log KOC . However, in our partitioning experiment,
we found no evidence of enhanced or reduced PAH
sorption relative to the bulk sediment in either of the
fractions examined. Observed OC-normalized parti-
tioning coefficients (log KOC ) for pyrene were 4.93, 4.88
and 4.94 for the <63 lm, 125–250 lm and bulk sediment
fractions, respectively, and closely matched the theo-
retical KOC value of 4.86 for generic OC, as calcu-
lated from pyrene’s mean log KOW of 5.07 (Meador
et al., 1995) and assuming the relationship log KOC ¼
log KOW  0:21 (Karickhoff et al., 1979). The finding of
similar and typical sorption affinity for pyrene among
the different sediment fractions analyzed suggests similar
bioavailability of PAHs in the different size fractions of
MC sediment. Analogously, Ahrens and Hickey (2003)
found little evidence for enhanced fluoranthene sorption
in MC sediment, despite its heterogeneous OC compo-
sition. In light of these results, the hypothesis that PAHs
in MC sediment are associated with super-sorbent soot
phases cannot be maintained. However, this does not
rule out that other forms of anthropogenic BC, such as
coal tar could be the main PAH sorption phases. In fact,
the pyrogenic signature and specific distribution of al-
kylated phenanthrenes, in addition to the very high
concentrations of PAHs in MC sediment, strongly sug-
gest contamination by coal-tar or pitch (Depree and
Ahrens, 2003). While there is no apparent current source
349
of coal tar in Motions Creek, this material may repre-
sent a historic burden from outdated road-building
practices or manufactured gas production. Interestingly,
Ghosh et al. (2003) reported that coal-tar does not show
enhanced sorption of PAHs relative to generic OC, in
strong contrast to other forms of BC, such as soot, coal,
coke and activated carbon: a summary of OC-normal-
ized partition coefficients for phenanthrene shows the
binding affinity of coal tar (log KOC ¼ 5) to be two or-
ders of magnitude less than that of activated carbon
(log KOC ¼ 7).
4. Conclusions
Our finding that PAH concentrations in Motions
Creek varied by over 100-fold between size and density
sub-fractions of the same sediment sample may have
useful implications for environmental management. The
fact that more than 75% of PAHs were concentrated on
light, organic-rich particles that comprise only a small
fraction of the sediment mass means that sediment PAH
concentrations and inputs could be predicted to a large
degree by focusing on the transport and distribution of
only this concentrated sub-fraction within the catch-
ment. Furthermore, since this light material comprises
only 3% of the sediment mass, interception or post-
depositional removal of this light fraction may be an
efficient way to diminish the delivery of PAHs to benthic
biota. The same conclusion was reached by Rockne
et al. (2002), who suggested targeted removal of the
PAH-enriched low density fraction in sediments from
the New York/New Jersey Harbor. While we currently
have only limited information on the bioavailability of
PAHs in the OC-rich light fractions in Motions Creek, it
can be argued that the 10–200 fold higher PAH con-
centrations in low density fractions probably more than
compensate for any potential reductions in bioavail-
ability. Our partitioning comparison using pyrene sup-
ports the assumption that PAHs are similarly available
among different sediment fractions. Only if L-fraction
PAH bioavailability were decreased more than 10–200
fold relative to H-fractions, would removal of the light
fraction have no impact on biota.
Alternatively, if the PAHs found to be concentrated
in the light, OC-rich sediment fractions did not originate
on these particles but re-partitioned to these only after
their delivery to the estuary, the remarkably high RPAH
concentrations would suggest that PAHs are rapidly
scavenged by the OC of this fraction. If this is the case,
post-depositional removal of PAHs may be feasible by
deploying sediment OC surrogates, such as activated
carbon, which could be buried in contaminated sedi-
ments to ‘‘soak up’’ PAHs and removed after a specified
duration. Both scenarios––delivery of PAHs in a con-
centrated sediment fraction, or intensive re-partitioning
350 M.J. Ahrens, C.V. Depree / Marine Pollution Bulletin 48 (2004) 341–350
and concentration into a low-density OC fraction in
situ––make removal of high PAH concentrations a
tangible option in Motions Creek.
Acknowledgement
We gratefully acknowledge financial support from the
New Zealand Foundation for Research Science and
Technology.
References
Ahrens, M.J., Hickey, C.W., 2003. Extractability of polycyclic
aromatic hydrocarbons in sediments: A matter of association?
Australian Journal of Chemistry 56, 193–199.
Burgess, R.M., Ahrens, M.J., Hickey, C.W., den Besten, P.J., ten
Hulscher, D., van Hattum, B., Meador, J.P., Douben, P.E.T., 2003.
An overview of the partitioning and bioavailability of PAHs in
sediments and soils. In: Douben, P. (Ed.), PAHs: An Ecotoxico-
logical Perspective. Wiley, Chichester, pp. 99–126.
Cornelissen, G., Rigterink, H., Ferdinandy, M.M.A., Van Noort,
P.C.M., 1998. Rapidly desorbing fractions of PAHs in contami-
nated sediments as a predictor of the extent of bioremediation.
Environmental Science & Technology 32, 966–970.
Cuypers, C., Grotenhuis, T., Joziasse, J., Rulkens, W., 2000. Rapid
persulfate oxidation predicts PAH bioavailability in soils and
sediments. Environmental Science & Technology 34, 2057–2063.
Depree, C.V., Ahrens, M.J., 2003. Distribution of polycyclic aromatic
hydrocarbons (PAHs) in an urbanised estuary and possible
implications for source apportionment. In: Tremblay, H., Locat,
J., Galvez-Cloutier, R., (Eds.), Second International Symposium
on Contaminated Sediments. Hotel Loews Le Concorde, Quebec,
Canada, pp. 182–188.
Di Toro, D.M., Zarba, C.S., et al., 1991. Technical basis for estimating
sediment quality criteria for nonionic organic chemicals using
equilibrium partitioning. Environmental Toxicology and Chemis-
try 10, 1541–1583.
Evans, K.M., Gill, R.A., Robotham, P.W.J., 1990. The PAH and
organic content of sediment particle size fractions. Water, Air, &
Soil Pollution 51, 13–31.
Fang, M.-D., Lee, C.-L., Yu, C.-S., 2003. Distribution and source
recognition of polycyclic aromatic hydrocarbons in the sediments
of Hsin-ta Harbour and adjacent coastal areas, Taiwan. Marine
Pollution Bulletin 46, 941–953.
Ghosh, U., Gillette, J.S., Luthy, R.G., Zare, R.N., 2000. Microscale
location, characterization, and association of polycyclic aromatic
hydrocarbons on harbor sediment particles. Environmental Science
& Technology 34, 1729–1736.
Ghosh, U., Zimmerman, J.R., Luthy, R.G., 2003. PCB and PAH
speciation among particle types in contaminated harbor sediments
and effects on PAH bioavailability. Environmental Science &
Technology 37, 2209–2217.
Gustafsson, O., Gschwend, P.M., 1997. Soot as a strong partition
medium for polycyclic aromatic hydrocarbons in aquatic systems.
In: Eganhouse, R.P. (Ed.), Molecular Markers in Environmental
Geochemistry. American Chemical Society, Washington, DC, pp.
365–381.
Gustafsson, O., Haghseta, F., Chan, C., MacFarlane, J., Gschwend,
P.M., 1997. Quantification of the dilute sedimentary soot phase:
Implications for PAH speciation and bioavailability. Environmen-
tal Science & Technology 31, 203–209.
Hickey, C.W., Roper, D.S., Holland, P.T., Trower, T.M., 1995.
Accumulation of organic contaminants in two sediment-dwelling
shellfish with contrasting feeding modes: Deposit- (Macomona
liliana) and filter-feeding (Austrovenus stutchburyi). Archives of
Environmental Contamination and Toxicology 29, 221–231.
Hoffman, E.J., Mills, G.L., Latimer, J.S., Quinn, J.G., 1984. Urban
runoff as a source of polycyclic aromatic hydrocarbons to coastal
waters. Environmental Science & Technology 18, 580–587.
Huang, W., Young, T.M., Schlautman, M.A., Yu, H., Weber W.J. Jr.,
1997. A distributed reactivity model for sorption by soils and
sediments. 9. General isotherm nonlinearity and applicability of the
dual reactivity domain model. Environmental Science & Technol-
ogy 31, 1703–1710.
Karickhoff, S.W., Brown, D.S., Scott, T.A., 1979. Sorption of
hydrophobic pollutants on natural sediments. Water Research 13,
241–248.
Maruya, K.A., Risebrough, R.W., Horne, A.J., 1996. Partitioning of
polynuclear aromatic hydrocarbons between sediments from San
Francisco Bay and their porewaters. Environmental Science &
Technology 30, 2942–2947.
McCready, S., Slee, D.J., Birch, G.F., Taylor, S.E., 2000. The
distribution of polycyclic aromatic hydrocarbons in surficial
sediments of Sydney Harbour, Australia. Marine Pollution Bulletin
40, 999–1006.
McGroddy, S.E., Farrington, J.W., 1995. Sediment porewater parti-
tioning of polycyclic aromatic hydrocarbons in three cores from
Boston Harbor, Massachusetts. Environmental Science & Tech-
nology 29,, 1542–1550.
McGroddy, S.E., Farrington, J.W., Gschwend, P.M., 1996. Compar-
ison of the in situ and desorption sediment-water partitioning of
polycyclic aromatic hydrocarbons and polychlorinated biphenyls.
Environmental Science & Technology 30, 172–177.
Meador, J.P., Stein, J.E., Reichert, W.L., Varanasi, U., 1995.
Bioaccumulation of polycyclic aromatic hydrocarbons by marine
organisms. Reviews of Environmental Contamination and Toxi-
cology 143, 79–165.
Prahl, F.G., Carpenter, R., 1983. Polycyclic aromatic hydrocarbon
(PAH)-phase associations in Washington Coastal sediments. Geo-
chimica et Cosmochimica Acta 43, 1013–1023.
Rockne, K.J., Shor, L.M., Young, L.Y., Taghon, G.L., Kosson, D.S.,
2002. Distributed sequestration and release of PAHs in weathered
sediment: The role of sediment structure and organic carbon
properties. Environmental Science & Technology 36, 2636–2644.
Simpson, C.D., Harrington, C.F., Cullen, W.R., 1998. Polycyclic
aromatic hydrocarbon contamination in marine sediments near
Kitimat, British Columbia. Environmental Science & Technology
32, 3266–3272.
Wang, X.C., Zhang, Y.X., Chen, R.F., 2001. Distribution and
Partitioning of Polycyclic Aromatic Hydrocarbons (PAHs) in
Different Size Fractions in Sediments from Boston Harbor, United
States. Marine Pollution Bulletin 42, 1139–1149.