This document describes the development of a solid phase microextraction-gas chromatography (SPME-GC) method to analyze the compounds released from Eucalyptus camaldulensis (River Red Gum) leaves and track this dissolved organic carbon (DOC) in aquatic environments. The researchers optimized the SPME extraction conditions using standards representative of compounds likely present in River Red Gum leaf leachate. They then applied the optimized method to analyze artificial DOC solutions prepared from River Red Gum leaf leachate to demonstrate the potential of the method for fingerprinting and tracing natural DOC sources in rivers.

1. Analytica Chimica Acta 530 (2005) 325–333
A solid phase microextraction method to fingerprint dissolved
organic carbon released from Eucalyptus camaldulensis (Dehnh.)
(River Red Gum) leaves
Alek Zandera, Andrea G. Bishopb, Paul D. Prenzlerb,∗
a Johnstone Centre, School of Science and Technology, Locked Bag 588, Charles Sturt University, Wagga Wagga, NSW 2678, Australia
b Farrer Centre, School of Science and Technology, Locked Bag 588, Charles Sturt University, Wagga Wagga, NSW 2678, Australia
Received 9 June 2004; received in revised form 13 September 2004; accepted 13 September 2004
Available online 7 December 2004
Abstract
A solid phase microextraction-gas chromatography (SPME-GC) method was developed to trace natural sources of dissolved organic carbon
(DOC) in river systems. The effects of extraction time, temperature, salt concentration, rate of stirring, and silanisation of sampling container
were examined. The optimum extraction conditions using a polydimethylsiloxane (PDMS) SPME fibre were found to be extraction for 15 min
at 40 ◦
C, pH 2, from a saturated NaCl matrix with rapid stirring in a non-silanised vial. The method gave good results for a series of six
compounds representative of those likely to be present in dissolved organic carbon leached from River Red Gum leaves—cineole, terpineol,
thymol, myristic acid, methyl palmitate and methyl stearate. Artificial dissolved organic carbon solutions prepared from River Red Gum leaf
leachate were also examined and the effects of filtering and storage on the filtrate were noted. The method was demonstrated to have potential
to track the leachate in aquatic environment, indicated by the large number of compounds extracted from leachate solutions, and the broad
linear working ranges of extracted compounds.
© 2004 Elsevier B.V. All rights reserved.
Keywords: SPME; SPME-GC; Water analysis; Environmental analysis; DOC
1. Introduction
Dissolved organic carbon (DOC) is a complex mixture
of compounds with molecular weights ranging from tens of
Daltons to thousands of Daltons [1]. It is commonly defined
as that fraction of carbon that passes through a 0.45 ␮m filter,
thus distinguishing it from various types of particulate carbon
[2].RiverwaterDOCiscomposedofcompoundsfromanum-
ber of chemical classes, including monomeric amino acids
and sugars, and polymeric fulvic and humic acids—with hu-
mic substances comprising at least 40% of the dissolved car-
bon [3]. DOC has been extensively investigated because of
its range of significant environmental effects in aquatic sys-
tems: as an energy source [4], as a mediator of nutrient and
∗ Corresponding author. Tel.: +61 2 6933 2978.
E-mail address: pprenzler@csu.edu.au (P.D. Prenzler).
metal ion availability [3] and as a regulator of photosynthesis
through light attenuation [3,5] among others.
Three major conceptual models have been advanced to
describe the factors responsible for variation in carbon trans-
portandtransformationsinriverineecosystems,viz.theRiver
Continuum Concept [6], that stresses the importance of or-
ganic carbon supplied from upper catchment sources; the
Flood Pulse Concept [7], that stresses the importance of or-
ganic carbon from floodplains; and the Riverine Productivity
model [8], that stresses the importance of algal production
in situ. These models have been mostly based on work in
rivers with relatively predictable flows, unlike the majority
of rivers flowing through inland Australia [9]. To advance
the theoretical framework of river modelling, especially as it
applies to Australian rivers, a chemical marker that could dis-
tinguish sources of DOC – e.g. catchment versus floodplain
– is necessary.
0003-2670/$ – see front matter © 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.aca.2004.09.032

2. 326 A. Zander et al. / Analytica Chimica Acta 530 (2005) 325–333
For many Australian floodplain rivers, Eucalyptus camal-
dulensis (River Red Gum) is the most widely distributed tree
species found on riverbanks and floodplains [10]. It has the
potential to make a significant contribution to the total car-
bon entering a river system following floodplain inundation
[11,12]. Furthermore, it has been shown that DOC from Red
Gum is leached most readily from the leaves (as opposed to
bark, twigs, etc. [12]). Therefore a technique that could fin-
gerprint this DOC leachate and follow this fingerprint in a
water body would greatly assist development of models of
river function.
Solid phase microextraction (SPME) coupled with gas
chromatography (GC) is now established as a versatile tech-
nique to analyse volatile and semi-volatile compounds in
aqueous solution [13]. There are many advantages in using
SPME, but two that are most relevant for this work are the
minimal sample preparation required and the potential to de-
tect compounds down to very low levels (ppt have been re-
ported [14]). The first of these opens the possibility of finger-
printing DOC sources directly from the water sample without
filtration through 0.45 ␮m filters—a process that can be ex-
tremely tedious. The second of these advantages means that
SPME has the potential of “tracking” the chemical finger-
print from a DOC source as that fingerprint becomes diluted
during its passage down a river.
In this study we investigate the potential of SPME-GC
to analyse the volatile and semi-volatile compounds released
fromRedGumleavesthathavebeensoakedinwater,i.e.from
an artificial DOC solution. The approach presented is atypical
of most analyses in that we have no target analytes—our tar-
get is a set of SPME conditions that maximise the fingerprint
of Red Gum leachate. In the absence of definitive knowledge
as to what the specific suite of compounds might be, we have
first systematically optimised the SPME method using a se-
lection of six standards. These standards are representative of
the classes of compounds that may be present in the leachate
[15–17]: two terpenes, cineole and terpineol; one phenol, thy-
mol; one long-chain fatty acid, myristic acid; and two fatty
acid esters, methyl palmitate and methyl stearate. The opti-
mised method is then applied to the analysis Red Gum leaf
leachate. To our knowledge this is the first example of the
application of SPME-GC to the analysis of the composition
of DOC in water samples.
2. Experimental
2.1. Materials
1,8-Cineole (Ajax Chemicals), ␣-terpineol (98%, Merck-
schuchardt), thymol (BDH Chemicals), and myristic acid
(98%), methyl palmitate (95%) and methyl stearate (97%)
all from Lancaster Synthesis, were used without further pu-
rification. Nanopure water from a ‘Modulab Analytical’ re-
search grade water system was used for the preparation of all
aqueous standards, solutions and extraction matrices. Other
solvents,acetoneandmethanol,wereofHPLCpesticideanal-
ysis grade.
2.2. Instrumentation
GC analysis was performed using a Varian Star 3400CX
gas chromatograph equipped with FID. Nitrogen was
used as the carrier gas through a Zebron ZB5 column
(30 m × 0.25 mm × 0.25 ␮m) at a rate of 2.0 mL/min. The
injector port was used in splitless mode at 270 ◦C and the de-
tector temperature was 300 ◦C. Two temperature programs
were used for the column oven, one for the extraction op-
timisation procedures using standards and another for the
analysis of the artificial DOC samples. For the standards the
program was: initial 40 ◦C, hold 4 min, ramp to 130 ◦C at
30 ◦C/min, ramp to 300 ◦C at 25 ◦C/min, hold 5 min. For
the DOC the program was: initial 40 ◦C, hold 4 min, ramp
to 100 ◦C at 30 ◦C/min, ramp to 110 ◦C at 1 ◦C/min, ramp
to 200 ◦C at 6 ◦C/min, ramp to 300 ◦C at 20 ◦C/min, hold
3 min.
The concentration of dissolved organic carbon was mea-
sured using a Dohrmann DC-190 total organic carbon anal-
yser calibrated with potassium hydrogen phthalate.
2.3. SPME
Supelco® SPME fibres, 100 ␮m PDMS (57300-U), and
holders (57330-U) were conditioned in the GC injector port
at 250 ◦C for 1 h prior to initial use. All extractions were per-
formed in an incubator (Watson Victor Ltd. model 70) for
temperature control. Scintillation vials were used to hold the
extraction matrix (15.0 mL), which was pre-warmed to ex-
traction temperature. A fixed clamp ensured the same relative
position of the SPME fibre in the matrix each time. Vigor-
ous stirring (Hannah Instruments HI190M) was maintained
throughout the extraction procedure. After removal from the
extraction matrix, the exposed fibre was swirled in nanop-
ure water for ten seconds to remove excess NaCl. There was
no noticeable decrease in extraction efficiency from this step
while failure to do so resulted in much reduced fibre life.
As an added measure, the fibre was left exposed in nanopure
water each night to remove remaining NaCl deposits. After
each extraction, the SPME fibres were desorbed for 4 min in
the GC injector port at 270 ◦C. Preliminary work indicated
that some residual carryover may occur using 2 min desorp-
tion, but blank runs indicated no such carryover using 4 min
desorption.
2.4. Preparation of standards
Stock standards (1.0 mg/mL) of cineole, terpineol, thy-
mol and myristic acid were prepared using methanol, and of
methyl palmitate and methyl stearate using methanol:acetone
(9:1). A combined standard (0.10 mg/mL) was prepared from
these using methanol:acetone (9:1) and stored at −18 ◦C.
Working standards (0.010 mg/mL) were prepared each week

3. A. Zander et al. / Analytica Chimica Acta 530 (2005) 325–333 327
from the combined standard and were kept at 4 ◦C. For all ex-
tractions, the appropriate volume of combined standard was
added to the pre-warmed extraction matrix immediately prior
to extraction.
2.5. SPME method development
All extractions were performed in triplicate using a to-
tal volume of 15.0 mL of extraction matrix and 150 ␮L of
working standard. SPME was performed immediately after
an aliquot of working standard had been added to the extrac-
tion matrix. The extraction was optimised for temperature
(25, 30, 35 and 40 ◦C), time (5, 10, 15, 20 min), salt con-
centration (0, 11, 22, 33% NaCl), pH (2.0, 4.0, 6.3, 8.5) ad-
justed with HCl or NaOH, rate of stirring and silanisation of
vials.
Calibration curves for each of the standard compounds
were prepared by addition of different volumes – 10, 50,
100, 150, 500, 1000 ␮L – of the working standard to the pre-
warmed(40 ◦C)extractionmatrix(33%NaCl,pH2)followed
by SPME for 15 min at 40 ◦C with rapid stirring of the sample.
The extractions were performed in triplicate.
2.6. Preparation of leachate from Red Gum leaves
Senescent leaves from River Red Gum were collected
from trees growing on the Murrumbidgee River floodplain
near Wagga Wagga, NSW. These leaves were air dried for 7
days then ground to a fine powder using an IKA Labortechnik
A10 mill. Leaching experiments were performed in triplicate
in 10 mL of 2.5 mM NaN3 solution in the dark. The pres-
ence of NaN3 inhibits microbial activity [18] that has been
shown [19] to affect the chemical make up of leachate from
Red Gum leaf powder. Leaching time was investigated using
1.0 g of leaf powder in solution for 10, 100 and 1000 min. The
mass of leaf powder was also varied—0.01, 0.10, and 1.0 g
in 10 mL. The leachate contained approximately 100 mg of
DOC for every 1 g of leaf powder used. Samples prepared in
this manner will be referred to as “artificial DOC” or “leaf
leachate”, below.
2.7. Effect of filtration, storage and concentration upon
chromatographic profile of Red Gum leachate
For these experiments, 2.0 g of leaf powder was added
to 200 mL of 2.5 mM NaN3 solution. After standing in
the dark for 100 min the leachate was filtered (Whatman
No. 1 and 0.45 ␮m mixed cellulose ester) to give approxi-
mately 1100 mg/L DOC. Four 200 mL volumes of leachate
were prepared as above. Half of each volume was filtered
with the other half left unfiltered. From each of the eight
100 mL leachate samples, 150 ␮L was added to pre-warmed
(40 ◦C, 15 mL) saturated NaCl matrix, and immediately ex-
tracted for 15 min. The eight samples of Red Gum leachate
were then stored at ∼4 ◦C for 14 days and re-extracted as
above.
A 200 mL volume of filtered Red Gum leachate was pre-
pared as above. From this stock, saturated NaCl solutions
were prepared with a final concentration of 0.03, 0.10, 0.45,
0.91, 3.3, 6.7 and 10 g/L leaf powder, with three replicates
for each concentration. Solutions with a concentration of
0.91 g/L and less were prepared by addition of the appropri-
ate volume of leachate to pre-warmed (40 ◦C) saturated NaCl
solution (15 mL). The higher concentrations were prepared
by addition of 4.5 g of NaCl to undiluted leachate followed
by warming for 30 min. Extraction was performed for 15 min
at 40 ◦C either immediately after addition of the leachate or
the expiration of the warming period.
3. Results and discussion
Since River Red Gum is the dominant species on Aus-
tralian floodplain rivers, its potential contribution to to-
tal aquatic DOC has been studied by several authors
[12,18,20,21]. Leaching experiments [12] showed that up
to 8% of the dry mass could be solubilised within 29 days
(the majority of this in less than 10 days). However, it is not
known what contribution this DOC makes to the ecology of
the aquatic environment since it has so far been impossible to
trace this leachate. Among the compounds possibly leached
from Red Gum leaves are terpenes, phenols, fatty acids and
fatty acid esters. The six compounds chosen in this study are
representative of these various classes and provide a series
of compounds that differ considerably in chromatographic
retention time, polarity, functional groups, and molecular
weight. As will be discussed below, these compounds also
display quite different behaviour under the conditions chosen
to optimise the SPME method. Thus it was possible to gain
insight as to how various classes of compounds responded to
alterations in the SPME parameters and how future studies
may be designed to target a particular group of compounds.
3.1. Optimisation of extraction conditions
Optimisation of the SPME method for a system of inter-
est usually involves selecting a fibre appropriate for the com-
pounds under investigation, followed by systematic variation
of extraction time and temperature. Further parameters that
may influence extraction efficiency are the ionic strength,
pH, and rate of stirring of the matrix [14] and adsorption of
analytes to container walls. The effect of the latter may be
minimised by silanisation of the container walls [22]. Each
of these parameters was investigated in this study for a sin-
gle fibre type. All extractions were performed with a PDMS
fibre since it has been reported [23] that this fibre coating
minimises interactions between dissolved humic substances
and the fibre.
The effects of extraction time, salt concentration, and pH
on the extraction efficiency of the six representative com-
pounds are shown in Figs. 1–3, respectively. A marked dif-
ference in the responses of the two terpenes, cineole and ter-

4. 328 A. Zander et al. / Analytica Chimica Acta 530 (2005) 325–333
Fig. 1. Effect of extraction time on recovery of standards as indicated by GC area counts. SPME conditions 40 ◦C, 33% NaCl; standard concentration 100 ␮g/L.
Fig. 2. Effect of extraction matrix ionic strength on recovery of standards as indicated by GC area counts. SPME conditions 40 ◦C, extraction time 15 min;
standard concentration 100 ␮g/L.
Fig. 3. Effect of extraction matrix pH on recovery of standards as indicated by GC area counts. SPME conditions 40 ◦C, 33% NaCl, extraction time 15 min;
standard concentration 100 ␮g/L.

5. A. Zander et al. / Analytica Chimica Acta 530 (2005) 325–333 329
pineol, is evident in the figures. This is most likely due to dif-
ferent partition coefficients for the PDMS fibre as these com-
pounds show similar FID responses when directly injected in
the GC. There were no statistically significant differences be-
tween the responses of cineole, terpineol and thymol at 15 or
20 min extraction even though equilibrium was not reached.
Longer extraction times, up to one hour (data not shown),
led to significant enhancement (>double) of recovery of the
fatty acid methyl esters, but at the expense of the terpenes and
thymol. As a compromise, an extraction time of 15 min was
used in subsequent experiments, and since equilibrium was
not reached at 15 min, extraction time was controlled very
precisely.
Fig. 1 also illustrates two other general observations from
this system. Firstly, the response of myristic acid is highly
variable(largecoefficientsofvariance)andunpredictable(es-
pecially Fig. 2). This may have implications for the present
application of SPME in that if free fatty acids make a sig-
nificant contribution to the DOC fingerprint then the current
optimised method (see below) would need to be re-optimised
for this class of compounds. Note that lipids comprise less
than 3% of the total mass in senescent leaves [24]. The second
general feature of this system is the low responses for the fatty
acid methyl esters. Again, if this class of compounds proves
important for the study of DOC, and sensitivity becomes an
issue, then it may be necessary to develop a new method,
e.g. increase extraction time. It must be emphasised that in
the absence of a specific target analyte any extraction method
will necessarily be a compromise [25].
The effect of matrix ionic strength on extraction efficiency
is shown in Fig. 2. Cineole, terpineol and thymol all show sig-
nificantly improved recoveries as the sodium chloride con-
centration is increased. The recoveries of the fatty acid methyl
esters do not appear to be affected by salt concentration, while
that of myristic acid is variable, but reaches a minimum in
saturated NaCl solution. Bucholz and Pawliszyn [26] demon-
strated that for some phenols, recoveries could increase by
up to a factor of five in saturated NaCl solution over solu-
tions containing no added salt. However, no intermediate salt
concentrations were tested. Fig. 2 shows that for some ana-
lytes the effect of NaCl addition is an incremental increase
in recovery such that only a saturated NaCl matrix will max-
imise recovery. The decrease in myristic acid recovery as
NaCl concentration is increased is consistent with the fact
that increasing ionic strength will enhance ionisation of the
acid and hence, increase its affinity for the aqueous matrix
[26].
The pH of the extraction matrix is expected to have an
effect on extraction efficiency when the target analyte is able
to participate in protonation equilibria [26], whereby the neu-
tral form of the molecule partitions on the fibre to a greater
extent. Pawliszyn [27] recommends a solution pH two units
above/below the pKa of the target analyte to ensure >99%
is in the un-ionised form. Thus the results shown in Fig. 3
are surprising: myristic acid would be expected to show a
significantly increased affinity for the fibre at pH 2, but does
not. This result would indicate that protonation state is not
critical to partitioning of myristic acid to the fibre, contrary to
the finding above. Further work appears necessary to under-
stand the mechanism of long-chain fatty acids partitioning on
PDMS fibres. The neutral cineole shows increasing recovery
to lower pH despite not being able to participate in proto-
nation equilibria and thymol also shows enhanced recovery
at low pH despite being essentially fully protonated at pH 6
and below. As observed for ionic strength, above, the fatty
acid methyl esters show little dependence on the pH of the
extraction matrix.
Extraction temperature, rate of stirring, and silanisation
of the extraction vessel were also tested to determine their
effect on recovery of the six compounds. Temperature was
found to only affect the recovery of myristic acid—with 25 ◦C
being the optimum temperature. However, because the am-
bient temperature of the laboratory routinely exceeds 30 ◦C
in summer, it was decided that 40 ◦C would be the extraction
temperature for future analyses. Silanisation of the extrac-
tion vessel was found to have no significant effect on recov-
eries of compounds. Rapid stirring increased recoveries of
the fatty acid methyl esters with little influence on the other
compounds.
From the above results the optimised method
was—extraction for 15 min at 40 ◦C, pH 2, from a saturated
NaCl matrix in a non-silanised vial with rapid stirring. Rela-
tive errors for all compounds, excluding myristic acid, were
generally ≤10%. They were sometimes a little higher for
the fatty acid methyl esters, simply because recoveries were
lower and as noted above, higher recoveries could be gained
with longer extraction times. Using this method the cali-
bration curves for terpineol, myristic acid, methyl palmitate
and methyl stearate showed good linearity (r2 ≥ 0.98) in the
range 0–100 ␮g/L. The linear range for cineole and thymol
was 0–50 ␮g/L (r2 ≥ 0.98). Overall, the results for this repre-
sentative group of compounds demonstrate the potential for
this method in the analysis of real DOC samples (see below).
3.2. Fingerprinting leachate from Red Gum leaves
Initially, artificial DOC was prepared by suspending 1.0 g
of Red Gum leaf powder in 100 mL of H2O for 100 min at
room temperature ([DOC] ca. 1000 mg/L, see above). This
was followed by SPME of the filtrate using the optimised
conditions above; the gas chromatogram of the compounds
sorbed to the fibre is shown in Fig. 4. Using the temperature
program as described in Section 2, the peaks separate into two
regions, one with retention times of 8–18 min, and the second
with retention times of 26–33 min. From the chromatogram,
a total of 163 peaks were detected indicating that a myriad of
volatile and semi-volatile compounds are released from leaf
powder.
For any future quantitative work it was necessary to es-
tablish whether there is a linear relationship between mass
of leaf powder and recovery of compounds. Table 1 gives
the linear range, slope of the line of best fit and correlation

6. 330 A. Zander et al. / Analytica Chimica Acta 530 (2005) 325–333
Fig. 4. Representative gas chromatogram of artificial DOC (Red Gum leachate) prepared from 1.0 g leaf powder in 100 mL water for 100 min. SPME using
PDMS fibre, extraction time 15 min, extraction conditions 40 ◦C, 33% NaCl, pH 2.
coefficient for 33 peaks with area counts greater than 30,000
at 10 g/L leaf powder. Most of the peaks give good straight
lines (r2 > 0.98) up to 10 g/L leaf powder. The lower limit of
the linear range is set by the ability to integrate small peak
areas, especially in crowded regions of the chromatogram.
The 33 peaks have a range of slopes from 2000 to 750,000
counts per unit concentration leaf powder. This indicates that
the fibre has a wide range of responses to the compounds.
For those compounds that show very large slopes, the upper
limit of linearity is <10 g/L.
For future studies, where the SPME technique will be used
to fingerprint other potential sources of DOC (e.g. exotic tree
species and herbaceous plants), it was necessary to inves-
tigate the rate of release of the volatile/semi-volatile from
the leaf powder. Earlier work [12,19] has shown that when
a colourimetric measure of DOC is used (e.g. A440) DOC
release plateaus after 3 days. On the other hand, the types of
compounds sorbing to the PDMS fibre may be released much
more quickly than the polymeric material that gives rise to
the yellow “A440” colour that is typically correlated to DOC.
Fig. 5 illustrates the percentage change in peak area for 16
peaks (>50,000 area count) as release time is increased from
10 to 1000 min. Generally, peak areas increased by between
35 and 75% when the release time was increased from 10
to 100 min with a few notable exceptions where the increase
was greater than 200%. However, very few peaks showed
significant increase in concentration as the release time was
increased from 100 to 1000 min, while a few had peak areas
at 1000 min that were less than their peak areas at 10 min
release time.
It would appear from these results that the majority of
compounds that are released from Red Gum leaf powder
and amenable to SPME-GC are rapidly released, compared
to those that contribute to the yellow colour typically used
Fig. 5. DOC release from Red Gum leaf powder as a function of time. Plot of relative change in peak area (%) for peaks with area counts >50,000 after 10 min
extraction. Samples extracted after 10, 100, 1000 min with changes measured relative to area counts at 10 min. SPME using PDMS fibre, extraction time 10 min,
extraction conditions 40 ◦C, 33% NaCl, pH 2.

7. A. Zander et al. / Analytica Chimica Acta 530 (2005) 325–333 331
Fig. 6. Effects of filtration and storage on artificial DOC SPME-GC profile. Artificial DOC prepared from 1.0 g leaf powder in 100 mL water for 100 min.
SPME using PDMS fibre, extraction time 15 min, extraction conditions 40 ◦C, 33% NaCl, pH 2. (a) Effect of filtration: comparison of SPME-GC for a fresh
artificial DOC sample, filtered and unfiltered. (b) Effect of storage: comparison of SPME-GC for an unfiltered artificial DOC sample fresh and after 14 days
stored at 4 ◦C. (c) Effect of storage: comparison of SPME-GC for a filtered artificial DOC sample fresh and after 14 days stored at 4 ◦C.
as a measure of DOC. Further work, beyond the scope of
this study, will be needed to establish the ecological sig-
nificance of this difference in release rate from Red Gum
leaves.
Further studies were undertaken on the Red Gum leaf
leachate to investigate the effects of filtration and storage
on the recovery of released compounds. Filtration is a neces-
sary pre-requisite to determining DOC concentration, and has
the potential to alter the volatile/semi-volatile profile through
various mechanisms such as evaporation of volatile com-
pounds and removal of compounds on the particulate matter
or filter membranes. Fig. 6a indicates that the most volatile
compounds are particularly affected by filtration. Eight of
the first 10 eluting peaks show significant reductions in peak
areas following filtration. Four of the remaining peaks also
show decreased peak areas in the chromatogram after filtra-
tion. The results indicate that the SPME-GC method may
be applied to filtered solutions to obtain a qualitative finger-
print of the DOC, but that quantitative determinations may
be compromised by filtration.
Storage of natural waters may be unavoidable in situa-
tions where sampling takes place long distances from the
laboratory, or where the number of samples is large. The
above results would tend to indicate that filtration prior to
storage is undesirable, due to loss of compounds. On the
other hand, storage of samples containing particulate organic
matter may result in further leaching of substances and an in-
crease in concentration of some compounds. Fig. 6b indicates
that the latter indeed occurs at 4 ◦C over 14 days storage. By
far the majority of compounds show an increase in peak area,

8. 332 A. Zander et al. / Analytica Chimica Acta 530 (2005) 325–333
Table 1
RelationshipsbetweenmassofleafpowderandrecoveryofDOCcompounds
from artificial DOC, using SPME-GC
Retention time
(min)
Linear range
(g powder/L)
Slope
(×103 counts L/g powder)
R2
8.78 0.91–6.70 11.53 0.981
8.94 0.03–3.3a 748 0.996
9.68 0.45–10.0 12.28 0.992
10.37 0.03–10.0 58.73 0.999
10.44 0.10–10.0 19.62 0.991
10.94 0.10–10.0 11.69 0.996
11.07 0.10–10.0 32.22 0.995
11.59 0.10–10.0 37.73 0.997
11.75 0.45–10.0 8.92 0.993
12.86 0.03–10.0 141 0.997
13.09 0.45–10.0 20.15 0.992
13.21 0.03–10.0 280 0.996
13.32 0.03–10.0 54.67 0.996
13.97 0.91–10.0 10.12 0.990
15.38 0.03–10.0 187 0.997
15.79 0.45–10.0 31.33 0.994
16.04 0.45–10.0 16.47 0.994
17.02 0.10–10.0 33.43 0.990
17.63 0.45–10.0 39.22 0.979
18.01 0.45–10.0 14.63 0.997
18.72 0.91–10.0 1.69 0.982
19.35 0.91–10.0 12.07 0.983
26.60 0.03–3.3a 476 0.994
29.32 0.45–10.0 8.65 0.996
29.67 0.45–10.0 11.47 0.997
29.77 0.45–10.0 7.44 0.986
29.87 0.45–10.0 6.06 0.997
30.15 0.45–10.0 8.57 0.994
30.27 0.10–10.0 17.39 0.995
30.70 0.91–10.0 4.21 0.983
31.34 0.91–10.0 8.99 0.989
32.07 0.10–10.0 11.13 0.997
32.87 0.10–10.0 11.16 0.998
Data presented for GC peaks with area counts >30,000 at 10 g/L leaf powder.
a Curve flattens out at higher concentration of leaf powder.
presumably due to leaching during storage. While filtration
prior to storage will counter this, storage of filtered leachate
samples at 4 ◦C does not render the system inert (Fig. 6c).
Nearly one third of the peaks show some variation on stor-
age, with most tending to decrease. Ideally, samples should
be analysed as soon as possible after collection. Where this
is not possible, storage times should be kept constant to min-
imise variations in the GC fingerprint that may occur with
time.
4. Conclusions
This paper has illustrated the potential for SPME-GC
as a tool to fingerprint and track dissolved organic car-
bon in rivers. The potential has been shown both through
the large number of GC peaks generated from artificial
DOC samples, and through the linearity of concentration-
recovery data. High reproducibility has been demonstrated
using the method. Several parameters required optimisation
and careful control: matrix ionic strength and pH, extraction
time and temperature—with the optimised method consist-
ing of extraction for 15 min at 40 ◦C, pH 2, from a satu-
rated NaCl matrix with rapid stirring in a non-silanised vial.
Sample treatment and storage methods have been shown to
be important factors, especially for samples that may con-
tain particulate organic matter. Future work will establish
whether long-chain fatty acids and/or fatty acid methyl es-
ters are important compounds in the DOC fingerprint. If so,
longer extraction times would be needed to optimise for
these compounds and other issues to do with the variabil-
ity of recovery of long-chain fatty acids would need to be
addressed.
Acknowledgements
The authors wish to acknowledge Alistar Robertson,
Kevin Robards and Darren Ryder for their continued sup-
port and input, and Charles Sturt University for funding this
work.
References
[1] R.L. Malcolm, Anal. Chim. Acta 232 (1990) 19.
[2] R.S. Wotton (Ed.), The Biology of Particles in Aquatic Systems,
Lewis Publishers, Boca Raton, 1994.
[3] S. Findlay, R.L. Sinsabaugh, Mar. Freshw. Res. 50 (1999) 781.
[4] J.D. Thomas, Freshw. Biol. 38 (1997) 1.
[5] S.R. Carpenter, J.L. Cole, J.F. Kitchell, M.L. Pace, Limnol.
Oceanogr. 43 (1998) 73.
[6] R.L. Vannote, G.W. Minshall, K.W. Cummins, J.R. Sedell, C.E.
Cushing, Can. J. Fish. Aquat. Sci. 37 (1980) 130.
[7] W.J. Junk, P.B. Bayley, R.E. Sparks, in: D.P. Dodge (Ed.), Proceed-
ings of the International Large River Symposium, Can. Spec. Publ.
Fish. Aquat. Sci. 106 (1989) 110.
[8] J.H. Thorp, M.D. Delong, Oikos 96 (2002) 543.
[9] J.T. Puckridge, F. Sheldon, K.F. Walker, A.J. Boulton, Mar. Freshw.
Res. 49 (1998) 55.
[10] N.C.W. Beadle, The Vegetation of Australia, Cambridge University
Press, Cambridge, 1981.
[11] A.I. Robertson, S.E. Bunn, P.I. Boon, K.F. Walker, Mar. Freshw.
Res. 50 (1999) 813.
[12] M. O’Connell, D.S. Baldwin, A.I. Robertson, G. Rees, Freshw. Biol.
45 (2000) 333.
[13] J. Pawliszyn, Applications of Solid Phase Microextraction, The
Royal Society for Chemistry, Cambridge, 1999.
[14] A.A. Boyd-Boland, S. Magdic, J.B. Pawliszyn, Analyst 121 (1996)
929.
[15] C.J. Volk, C.B. Volk, L.A. Kaplan, Limnol. Oceanogr. 42 (1997)
39.
[16] D.J. Boland, J.J. Brophy, A.P.N. House, Eucalyptus Leaf Oils: Use,
Chemistry, Distillation and Marketing, Inkata Press, Melbourne,
1991.
[17] C.C. Parrish, M.T. Arts, B.C. Wainman, Lipids in Freshwater Ecosys-
tems, Springer-Verlag, New York, 1998, p. 4.
[18] D.S. Baldwin, Freshw. Biol. 41 (1999) 675.
[19] W. Jellett, Phenolic compounds as a potential indicator of the source
of organic carbon in flood-plain rivers, Unpublished Honours Thesis,
Charles Sturt University, Wagga Wagga, Australia, 2003.
[20] C. Francis, F. Sheldon, Hydrobiologia 481 (2002) 113.

9. A. Zander et al. / Analytica Chimica Acta 530 (2005) 325–333 333
[21] H.S. Glazebrook, A.I. Robertson, Aust. J. Ecol. 24 (1999) 625.
[22] L. Pan, M. Adams, J. Pawliszyn, Anal. Chem. 67 (1995)
4396.
[23] F.-D. Kopinke, J. Porschmann, A. Georgi, in: J. Pawliszyn (Ed.), Ap-
plications of Solid Phase Microextraction, Royal Society of Chem-
istry, Cambridge, 1999, p. 111.
[24] G.L. Mills, J.V. McArthur, C. Wolfe, J.M. Aho, R.B. Rader, Arch.
Hydrobiol. 152 (2001) 315.
[25] K. Robards, J. Chromatogr. A 1000 (2003) 657–691.
[26] K.D. Bucholz, J. Pawliszyn, Anal. Chem. 66 (1994) 160.
[27] J. Pawliszyn (Ed.), Applications of Solid Phase Microextraction,
Royal Society of Chemistry, Cambridge, 1999. p. 3.