Water Research 39 (2005) 38493862

Coagulation of humic substances and dissolved organic matter

with a ferric salt: An electron energy loss spectroscopy
investigation
A.-V. Jung
a,
, V. Chanudet
a
, J. Ghanbaja
b
, B.S. Lartiges
a
, J.-L. Bersillon
a
a
Laboratoire Environnement et Mineralurgie (LEM), UMR 7569 CNRS-INPL-ENSG, 15 avenue du Charmois, BP 40,
54501 Vanduvre, France
b
Service commnun danalyses (MET), Universite Nancy I, BP 239, 54 506 Vanduvre, France
Received 5 February 2005; received in revised form 7 July 2005; accepted 11 July 2005
Available online 19 August 2005
Abstract
Transmission electron microscopy (TEM) coupled with electron energy loss spectroscopy (EELS) and energy
dispersive X-ray spectroscopy (EDXS) was used to investigate the coagulation of natural organic matter with a ferric
salt. Jar-test experiments were rst conducted with a reconstituted water containing either synthetic or natural extracts
of humic substances, and then with a raw water from Moselle River (France). The characterization of the freeze-dried
coagulated sediment by EELS in the 250450 eV range, showed that Fe-coagulant species predominantly associate with
the carboxylic groups of organic matter, and that this interaction is accompanied by a release of previously complexed
calcium ions. The variation of Fe/C elemental ratio with iron concentration provides insightful information into the
coagulation mechanism of humic substances. At acid pH, Fe/C remains close to 3 over the whole range of iron
concentrations investigated, while a much lower atomic ratio is expected from the value of optimal coagulant dosage.
This suggests that a charge neutralization/complexation mechanism is responsible for the removal of humic colloids, the
aggregates being formed with both iron-coagulated and proton-neutralized organic compounds. At pH 8, the decrease
in Fe/C around optimal coagulant concentration is interpreted as a bridging of stretched humic macromolecules by Fe-
hydrolyzed species. Aggregation would then result from a competition between reconformation of humic chains around
coagulant species and collision of destabilized humic material. EELS also enabled a ngerpriting of natural organic
substances contained in the iron-coagulated surface water, N/C elemental analyses revealing that humic colloids are
removed prior to proteinic compounds.
r 2005 Elsevier Ltd. All rights reserved.
Keywords: Coagulation; Humic substances; Natural organic matter; Electron energy loss spectroscopy (EELS)
1. Introduction
Dissolved organic matter (DOM) represents the
heterogeneous mixture of organic colloids (excluding
living organisms) and truly dissolved organic molecules
present in aquatic systems. It comprises biochemically
dened compounds such as proteins and carbohydrates,
ARTICLE IN PRESS
www.elsevier.com/locate/watres
0043-1354/$ - see front matter r 2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.watres.2005.07.008

Corresponding author. Private domicile, 84 chemin du

Klettenberg, 68350 Brunstatt, Brunstatt, France.
Tel.: +0033 6 32464105; fax: +0033 3 89647787.
E-mail address: av1jung@hotmail.com (A.-V. Jung).
operationally dened substances such as fulvic and
humic acids, and any supramolecular assembly of those
molecules and macromolecules (Leenheer and Croue ,
2003). DOM has largely been studied in recent years, as
it may play a major role in the mobility and bio-
availability of anthropic contaminants in the environ-
ment (Thurman and Malcolm, 1981; Tipping, 1993). In
drinking water treatment, the main motivation for
studying DOM stems from its role as precursor of
carcinogenic chlorination by-products during the disin-
fection stage (Oliver and Visser, 1980). However, most
problems encountered in the treatment process can be
related to the presence of DOM, including reduced
adsorption capacity in activated carbon beds, taste and
odour of the nished water, and biological regrowth in
the distribution networks (Owen et al., 1995).
The removal of DOM from water supplies is carried
out in the rst place by coagulation-occulation with a
hydrolyzable metal salt such as aluminum or iron
(Dempsey et al., 1984; Edwards and Amirtharajah,
1985; Gregor et al., 1997). When added to water,
aluminum and iron coagulants yield hydrolysis products
that destabilize the organic material present in the raw
water. The destabilization mechanisms most often
quoted in the literature include charge neutralization,
enmeshment, and adsorption (Dempsey et al., 1984).
Charge neutralization preferentially applies at acidic
pH, while sweep-occulation and adsorption are com-
monly referred to in pH-coagulant concentration
domains where abundant precipitation of metal hydro-
xide is expected (Edwards and Amirtharajah, 1985).
However, DOM-specic characteristics, such as its
molecular weight distribution, carboxylic acidity, and
humic substance content, have been shown to signi-
cantly inuence the efciency of DOM removal (Collins
et al., 1985; Bose and Reckhow, 1998; Matilainen et al.,
2002). Recent extended X-ray absorption ne spectro-
scopy (EXAFS) at the Fe K-edge experiments also
questioned the coagulation mechanisms usually invoked
in the literature, as poorly polymerized Fe species were
evidenced in a wide range of operational conditions
(Vilge -Ritter et al., 1999a, b; El Samrani et al., 2004).
As an optimized DOM coagulation is a key to better
drinking water treatment, the interaction behavior
between organic compounds and coagulant species
should be better understood. The approach taken in
this work was to investigate, within ocs, both the
nature of DOM and the speciation of an iron-based
coagulant by using electron energy loss spectroscopy
(EELS) combined with energy dispersive X-ray spectro-
scopy (EDXS). EELS provides a quantitative analysis at
the nanometer scale of light elements (carbon, nitrogen,
oxygen) through the K-shell excitation, and of other
important elements in water treatment such as iron,
calcium, phosphorus, and sulfur, through the L-shell
excitation. It also enables the identication of organic
chemical bonds within polymeric systems (Varlot et al.,
1998), and can then be used to ngerprint the various
organic macromolecules contained in coagulated organ-
ic matter. In environmental science, EELS was rst used
by biochemists to describe in situ natural materials
degradation with C/N ratio measurements (Villemin et
al., 1995; Watteau et al., 1996). It has also been used to
investigate the association of organic matter with clay
particles (Furukawa, 2000).
In this paper, we present the results obtained by C-
EELS. Fe-EELS results will be reported in a companion
paper (Jung et al., 2005a). To avoid problems associated
with changes in the quality of the river water, experi-
ments were rst carried out with a reconstituted river
water containing either humic-like synthetic compounds
or natural aquagenic humic extracts. In this way, the
inuence of pH on humics removal could be investi-
gated. The method was then assessed on raw Moselle
river water.
2. Experimental section
2.1. Origin of natural organic matter (NOM) and
extraction of humic substances
Three sources of organic matter were selected for this
study. A synthetic humic-like substance (SHS) was
prepared by polymerization of catechol with glycine
(Andreux et al., 1980; Jung et al., 2005b). Catechol
(6.6 g) was dissolved in 500 mL of a solution of sodium
phosphate buffer NaH
2
PO
4
/Na
2
HPO
4
at pH 8 to obtain
a 0.03 M catechol solution. This was then mixed with an
equal volume of buffered glycine 0.03 M solution. The
resulting mixture was then allowed to react in a 2 L ask
in the dark at 25 1C and under a pure oxygen constant
pressure (0.9 bar) for 5 days. After dialysis, (Spectra/
Por
s
7 membrane- 8 kDa molecular weight cut-off
Roth-Sochiel Lauterbourg, France), the synthetic mate-
rial in a sodium form was freeze-dried and stored at 4 1C
in the dark. Extensive characterization of SHS and
relevance to natural humic substances have been
discussed in a previous work (Jung et al., 2005b).
Elemental analysis gave a carbon to nitrogen ratio (C/N)
of 18.3 (%C w/w 40.7). Potentiometric titration
showed three acidities at pK
A
4.6 (2.3 meq/g), 5.1
(0.8 meq/g) and 7.2 (0.6 meq/g).
A natural aquagenic humic substance, hereafter
referred to as NHS, was extracted from the Moselle
river suspended matter. The sampling site was located at
Richardme nil (France) in the vicinity of the pumping
facility supplying Grand Nancy Urban Community
drinking water treatment plant in surface water.
Approximately, 12 m
3
of river water were pumped
(1 m
3
/h) into a ow-through centrifuge operating at
17,000 rpm. The humic substances contained in the
ARTICLE IN PRESS
A.-V. Jung et al. / Water Research 39 (2005) 38493862 3850
suspended material were then extracted following the
procedure of Thurman and Malcolm (1981), adapted to
yield the natural extract in a sodium form. NHS was
also puried by dialysis (Spectra/Por
s
7 membrane
8 kDa molecular weight cut-off) and freeze-dried.
Total exchangeable protons and C/N ratios determined
for NHS were 3.0 meq/g and 9.2 (%C w/w 44.2),
respectively.
Grab samples of the Moselle river water (9.3 mg/L
total organic carbon content) were also taken at the
sampling site, thus providing a raw water with
unfractionated NOM.
2.2. Coagulation procedure
Synthetic waters were prepared by dissolving 30 mg of
humic material (either SHS or NHS) in 1 L of Milli-Q
s
ultra-pure water. 336 mg of NaHCO
3
(4 10
3
mol/L)
(Sigma-Aldrichanalytical grade) was also added to the
suspension to provide a carbonate alkalinity similar to
that of natural river waters. Prior to coagulant injection,
the pH of synthetic waters was adjusted to pH 6 or 8
with dropwise addition of 0.1 M HCl. The surface water
was used as collected (natural pH 7.4). Jar-tests were
performed in 1 L bafed reactors of known power
dissipation characteristics (Lartiges et al., 1997). Suspen-
sion was stirred at 100 rpm which corresponds to a mean
velocity gradient of 135 s
1
. The coagulant, a 0.1 M
ferric nitrate solution prepared from crystallized
Fe(NO
3
)
3
, 9 H
2
O (95% pureSigma-Aldrich), was
injected under agitation using a 5000 mL micro-pipette
(Eppendorf, Hamburg, Germany), and mixing was
continued for 20 min. The suspension pH was not re-
adjusted after coagulant addition. After 24 h of quies-
cent settling in graduated Imhoff cones, the residual
turbidity of the supernatant was measured on a HACH
Ratio XR turbidimeter. The settled volume was
recorded, and the sediment was then collected and
freeze-dried (48 hEL105 Sentry) to perform TEM
observations, EDXS and EELS analysis.
2.3. Transmission electron microscopy (TEM)
Electron microscopy observations were performed
with a CM20 Philips transmission electronic microscope
(TEM) using a 200 kV accelerating voltage. The electron
microscope is coupled with an EDAX EDXS, and with a
GATAN 666 EELS. The freeze-dried sediment was re-
suspended in ethanol (99.5% spectroscopy grade) under
ultrasonication, and a drop of suspension was evapo-
rated on a carbon network-like holey support lm
placed on a 200 mesh copper grid (Euromedex
Mundolsheim, France). EDX spectra were recorded
with a 10 nm spot diameter and a 30 s counting time.
EELS spectra were collected in the 250450 eV region
with the same spot size (analyzed area of about 80 nm
2
).
The spectrometer entrance aperture, the half-angle of
collection and the dispersion were 2 mm, 10 mrad, and
0.2 eV/channel, respectively. The energy resolution,
determined at the full-width at half-maximum of the
zero-loss peak is 1.4 eV. The time acquisition was 2.5 s
for the low energy region and 8 s for the C K-edge
region. The relative thickness (ratio of sample thickness
to the free electronic inelastic diffusion pathway), was
kept about 0.570.2 for all measurements. The electron
dose, calculated from the Egerton relationship (Egerton,
1996), was about 101.9 C/m
2
which is less than the
critical value-inducing beam damage within organic
polymers (Varlot et al., 1998).
For each sample, 15 EEL spectra were recorded. All
data processing was carried out with GATAN EL/P 2.1
software. The spectra were rst corrected for dark
current, gain variations, and double excitation by
applying a Fourier-ratio deconvolution with weak loss
energy spectra. Background tting was performed with a
power law in order to extract the core-excitation peaks
for each element. The amounts of C, Ca, and N were
then obtained by integrating the surface area of a 50 eV
energy window after having dened the lower-limit in
energy for each element, that is 283.8, 346.4 and
401.6 eV for carbon (C K-), calcium (Ca L
2,3
-) and
nitrogen (N K-) edges, respectively. Integration results
are given by the software in atomic concentration/nm
2
.
Hence, Ca/C and N/C atomic ratios were directly
calculated from EELS measurements, whereas Na/C
and Fe/C ratios were inferred from the amounts in Na,
Ca, and Fe, provided by EDXS microanalyses. When
Ca was barely detectable on EEL spectra, which is the
case for most coagulation experiments carried out at pH
6, Fe/C and Na/C ratios obtained from EDXS spectra
were corrected by a proportional constant determined
from EELS and EDXS results at pH 8.
3. Results
3.1. Jar-test results
Fig. 1 presents the results obtained from jar-tests
carried out at pH 6 and 8 with SHS and NHS. All
turbidity and sediment volume curves display a classical
behavior: the residual turbidity steadily increases at low
coagulant concentration to reach a maximum, and then
steeply decreases upon further addition of ferric
coagulant. The point where the extrapolated steep
portion of the turbidity curve intersects the x-axis is
selected as the optimum coagulant concentration
(OCC), and is indicated with an arrow on the graphs.
At higher iron concentrations, the residual turbidity
remains low and then strongly increases as the
restabilization of the suspension is induced. The inverse
pattern is observed for sediment volume: at low iron
ARTICLE IN PRESS
A.-V. Jung et al. / Water Research 39 (2005) 38493862 3851
concentration, no settleable aggregate is observed up to
the maximum in supernatant turbidity. The sediment
volume then rapidly builds up just before the OCC, and
it levels off with further addition in coagulant. At pH 6,
the sediment volume gradually diminishes as the humic
suspension is restabilized.
OCC values, expressed in molar concentrations, and
pHs at OCC are given in Table 1. In all cases, the pH at
OCC falls within the 4.56 pH range for best removal of
organic substances (Dempsey et al., 1984). Surprisingly,
OCCs obtained for NHS coagulated under acid and
alkaline conditions are equivalent. As fresh stock
solutions of iron coagulant were prepared for each
experiment, this result could be due to the proteina-
ceaous nature of NHS (Jung et al., 2005b).
3.2. Morphology of coagulated humic substances
Typical micrographs of coagulated humic substances
are shown in Fig. 2. The sample preparation used in this
study precludes artifact free observations. Indeed,
freeze-drying is known to drastically modify the
structure of aggregated colloids (Hung et al., 1996),
whereas redispersion in ethanol may induce secondary
ARTICLE IN PRESS
Iron concentration (M)
0
2
4
6
8
10
12
14
16
18
0
5
10
15
20
25
0
2
4
6
8
10
12
0
10
20
30
40
50
60
0 2 10
-4
4 10
-4
6 10
-4
8 10
-4
10
-3
SHS pH 8
SHS pH 6
NHS pH 8 NHS pH 6
(a)
(b)
(c)
(d)
(e)
(f)
(g)
(h)
(i)
(j)
(k)
(l)
(m) ( n)
(o)
(p)
(q)
Iron concentration (M)
0 2 10
-4
4 10
-4
6 10
-4
8 10
-4
Iron concentration (M)
0 2 10
-4
4 10
-4
6 10
-4
8 10
-4
10
-3
Iron concentration (M)
0 4 10
-4
8 10
-4
1.2 10
-3
Fig. 1. Evolution of residual turbidity (K) and sediment volume () with coagulant concentration for NHS and SHS at pHs 6 and 8.
Arrows indicate optimal coagulant concentrations. The letters refer to freeze-dried sediment examined by TEM.
A.-V. Jung et al. / Water Research 39 (2005) 38493862 3852
agglomeration effects upon solvent removal (Stevenson
and Schnitzer, 1982). Actually, techniques such as
atomic force microscopy in aqueous solution (Maurice
and NamjesnikDejanovic, 1999; Plaschke et al.,
1999) or TEM observation of ultracentrifuged
and resin-coated colloids (Wilkinson et al., 1999),
seem more appropriate to provide a description of
hydrated humic material. Nevertheless, as the prepara-
tion conditions remained identical for all samples, the
morphology of coagulated humic substances could be
compared.
Coagulated SHS appears as compact aggregates of
overlapping spheroids about 30 nm in size, regardless of
pH and iron concentration (Figs. 2a and b). The
appearance of NHS aggregated at pH 8 is similar with
smaller spheroids (1015 nm) (Fig. 2c). In contrast,
electron micrographs of NHS coagulated at pH 6 reveal
a network of lament or sheet-like structures at the
micron scale (Fig. 2d). No signicant change of the
latter morphology could be detected with coagulant
concentration. Despite the structural changes certainly
induced by the preparatory technique, our TEM
observations evidence individual spheroids which aver-
age dimension is consistent with the range of sizes
previously reported in the literature for humic colloids
(Wilkinson et al., 1999; Plaschke et al., 1999).
3.3. EELS spectra of coagulated SHS and NHS
To the authors knowledge, EELS investigations of
NOM remain relatively scarce in the literature (Watteau
et al., 1996; Furukawa, 2000). The spectra of well-
characterized compounds such as Bovine serum albumin
and Dextran (polysaccharideM
W
100; 000) were
then recorded as standards, and the identication of
EELS spectral features was conducted in agreement with
recent near-edge X-ray absorption spectroscopy (NEX-
AFS) studies (Myneni, 2002 and references herein).
EELS spectra were calibrated in energy using the
346.4 eV Ca L
2,3
-edge as an internal reference. This
yielded a 401.6 eV for the N K-edge energy which is
consistent with the attribution usually proposed for
nitrogen in NOM (Watteau et al., 1996). When calcium
was not detected in our samples, the nitrogen K-edge
was used instead.
An example of energy loss spectrum recorded in the
250450 eV region (SHS coagulated at pH 8 with
[Fe] 7 10
4
mol/L), is illustrated in Fig. 3. Three
main features can be observed at the carbon edge: the
well-dened shoulder at 285 eV can be assigned to the
1s-p* transition of aromatic carbons (CQC); the
small peak around 288 eV is attributed to the 1s-p*
transition of the CQO bond in carboxylic groups, while
the broad peak centered at 300 eV is attributed to
overlapping s* transitions for CQC, CC, CH,
CQO, and CO bonds (Myneni, 2002). At higher
energy, the two peaks observed at 346 and 358 eV
correspond to the scattered electrons of the L23 shell of
calcium (Watteau et al., 1996). The presence of calcium
corresponds to a contamination brought about by the
addition of NaHCO
3
([Ca
2+
]2.7 10
4
mol/L). It is
likely associated with the functional groups of SHS as
calcium has been shown to compete with trace elements
for the binding sites of humic substances (e.g., Tipping,
1993; van den Hoop et al., 1995). Finally, in accordance
with the elemental composition of SHS, a small amount
of nitrogen can be detected at the N K-edge around
401 eV.
Increasing the coagulant concentration at pH 8
signicantly modies the electron loss spectrum of
coagulated SHS (Fig. 4a). At the carbon edge, the sharp
peak at 288 eV assigned to p* (CQO) drastically
decreases, while the large peak at 300 eV broadens
further. In addition, the peak at the calcium edge rapidly
drops. A recent NEXAFS study of pedogenic humic
acid also showed a similar decrease of the p* resonance
of the carboxyl group double bond upon aggregation
with Fe
3+
at pH 2 (Thieme et al., 2002). Such
observations can be readily explained by an exchange
of Fe-coagulant species with previously bonded Ca
2+
ions onto carboxylic groups of SHS. The same interac-
tion mechanism has also been inferred from infrared
spectroscopy results in the case of adsorption/desorption
of NOM on iron oxide particles (Gu et al., 1994),
coagulation of sewage with ferric chloride (El Samrani et
al., 2004), and more generally, such a reaction scheme is
consistent with the strong inuence of carboxylic group
ARTICLE IN PRESS
Table 1
Optimum coagulant (iron) concentrations (OCC) determined by jar-test procedures
OCC pH 8.0 pH 6.0 pH 7.4
SHS NHS SHS NHS NOM
Molar concentration (mol/L) 8.3 10
4
6.4 10
4
2.2 10
4
8.1 10
5
1.7 10
4
Reduced concentration (mg Fe/mg C
org
) 3.83 2.69 1.01 0.34 1.02
pH at the respective OCC 6.5 6.3 5.1 5.2 5.9
A.-V. Jung et al. / Water Research 39 (2005) 38493862 3853
content on coagulant demand (Lefebvre and Legube,
1990; Randtke, 1993; Collins et al., 1986; Specht et al.,
2000).
The inset of Fig. 4a presents the evolution of EELS
spectra of freeze-dried humic material remaining in the
supernatant after settling. For each iron concentration,
the spectrum closely resembles that of a sediment
aggregated with a lower coagulant dosage, and can be
characterized by a higher calcium content and by a
slightly sharper peak at 288 eV (p*CQO). This suggests
ARTICLE IN PRESS
Fig. 2. Electron micrographs of freeze-dried coagulated sediment: (a) SHS pH 8[Fe] 8.0 10
4
mol/L, (b) SHS pH 6
[Fe] 2.0 10
4
mol/L, (c) NHS pH 8[Fe] 7.0 10
4
mol/L, (d) NHS pH 6[Fe] 8.0 10
4
mol/L.
A.-V. Jung et al. / Water Research 39 (2005) 38493862 3854
that the unsettled humic material contains a lesser
amount of hydrolyzed Fe species. The EELS spectra of
NHS coagulated at pH 8 as a function of coagulant
concentration exhibit similar features (Fig. 4c): however,
the peak characterizing the 1s-p* transition of the
CQO bond in carboxylic groups is much less resolved
at low iron concentration, and it is gradually replaced by
a weak shoulder at higher coagulant dosage. Similarly,
the two peaks at the Ca L
2,3
edge, clearly dened at low
coagulant concentration, rapidly weaken with an
increase in iron concentration.
The EELS spectra of SHS and NHS coagulated at pH
6 as a function of iron concentration are shown in Figs.
4b and d. In that case, the relative intensities of the
various contributions do not vary noticeably with
coagulant concentration. At the carbon edge, the 1s-p*
transition of aromatic carbons (CQC) at 285 eV
remains as a distinct shoulder, whereas the contribution
at 288 eV (p* CQO) is now present as a weakly
resolved shoulder. The large dominant feature corre-
sponding to the s* transitions for CQC, CC, CH,
CQO, and CO bonds has shifted around 297 eV.
Finally, a large contribution in the 325 eV region can be
evidenced, but could not be assigned. However, the
freeze-dried humic material from the supernatant yield
EELS spectra similar to those obtained at pH 8 (inset of
Fig. 4b). In particular, the p* resonance of the CQO
double bond appears as a peak at low iron concentra-
tion, and this peak decreases at higher coagulant dosage.
Likewise, the Ca edge which was virtually absent in the
coagulated sediment, can be clearly observed except at
the highest iron concentration. Such observations
suggest that the residual turbidity is formed from a less
coagulated humic material also at pH 6.
3.4. Elemental analysis of coagulated SHS and NHS
Fig. 5 shows the evolution of N/C, Ca/C, Na/C, and
Fe/C atomic ratios of coagulated SHS and NHS with
iron concentration at pHs 6 and 8. N/C remains
constant as a function of coagulant concentration and
pH, thus indicating that both the extraction procedure
and the synthesis yielded a homogeneous humic material
for coagulation experiments. N/C atomic ratios calcu-
lated from EELS spectra are equal to 0.0770.02 and
0.1770.02 for SHS and NHS, respectively. Such values
are consistent with previous N/C elemental ratios
0.054 and 0.167obtained from chemical analysis (Jung
et al., 2005b). Similar trends are also observed for the
evolution of Ca/C and Na/C. At pH 8, Ca/C markedly
decreases just before OCC to reach an almost constant
elemental ratio at higher coagulant concentration,
ARTICLE IN PRESS
N
o
r
m
a
l
i
z
e
d

i
n
t
e
n
s
i
t
y

(
A
.
U
.
)
270 320 370 420
Energy Loss (eV)
285 eV
1s
C=C
288 eV
1s
C=C
296 - 297 eV
1s
C-N, C=O, C-O
C=C, C-C, C-H
C-K
D
e
r
i
v
a
t
i
v
e

o
f

n
o
r
m
a
l
i
z
e
d

i
n
t
e
n
s
i
t
i
e
s

(
A
.
U
.
)
293-294
299-300
286
285 288
297
275 310 345 380 415
C EELS
binding energies Energy Loss (eV)
Ca-L
2,3
N-K
Standard 1
Polysaccharide
Standard 2
Bovine serum
albumin
346.4
SHS
401.6
Ca L
2,3
-
N-K
Fig. 3. Typical EELS spectrum (SHS pH 8[Fe] 8.0 10
4
mol/L). The inset shows the double derivative for (i) Bovine serum
albumin, (ii) Polysaccharide, (iii) SHS at pH 8 and at [Fe] 8.0 10
4
mol/L.
A.-V. Jung et al. / Water Research 39 (2005) 38493862 3855
whereas Na/C remains close to 0.1 for both SHS and
NHS. At pH 6, Ca is not detected within the coagulated
sediment, whereas Na/C continuously decreases with
iron concentration in accordance with an exchange of
sodium ion for hydrolyzed Fe species onto anionic
groups of humic substances. The lower Na/C ratios
obtained at pH 8 are also consistent with an increased
deprotonation of SHS and NHS functional groups,
even though Na/C values may be modied during
freeze-drying.
The evolution of Fe/C atomic ratios as a function
of coagulant concentration follows two different
patterns according to pH. At pH 8, for both SHS
and NHS, Fe/C is about 3 before OCC, it surprisingly
drops to a value close to 1 at OCC (0.8 (c) and 1.25 (m)),
and then re-increases to an intermediate value at
higher coagulant dosage. At pH 6, Fe/C slightly
increases with iron concentration from a value of
1 at underdosage for NHS, to reach a Fe/C ratio
of about 3 at OCC for both humic compounds, and
ARTICLE IN PRESS
Fig. 4. EEL spectra of freeze-dried coagulated humic substances as a function of iron concentration: (a) SHS pH 8, (b) SHS pH 6,
(c) NHS pH 8, (d) NHS pH 6.
A.-V. Jung et al. / Water Research 39 (2005) 38493862 3856
values greater than 4 when the suspension is restabilized.
Interestingly, recent Fe K-edge EXAFS measure-
ments carried out on Seine river NOM coagulated
with iron chloride (Vilge Ritter et al., 1999a, b), revealed
that (i) iron hydrolysis is hindered to at most the
trimer stage, (ii) each iron atom is surrounded at
optimum dosage by three C and 1 C at pH 5.5 and
pH 7.5, respectively. Such results would yield Fe/C
elemental ratios in the 13 range which is consistent with
our data.
Assuming a complete removal of humic material by
coagulation, the determination of elemental ratios from
chemical analysis predicts that, at pH 8, Fe/C should be
equal to 0.8 and 0.66 for SHS and NHS, respectively,
which is in relatively good agreement with the values
obtained from EELS and EDX spectra. In contrast, at
pH 6, the same calculations yield Fe/C atomic ratios of
0.22 and 0.7 which is much less than the measured
values. Such a discrepancy may be explained by the
aggregation behavior of humic substances with pH. The
hydrolysis of iron coagulant implies a decrease in pH;
hence, at OCC, the nal pH of the suspension initially
coagulated at pH 6, was about 4.5. At the latter pH, a
small degree of aggregation of humic substances has
been clearly demonstrated from the measurement of
their diffusion coefcients by uorescence correlation
spectroscopy (Lead et al., 2000). In situ AFM observa-
tions of adsorbed humic colloids as a function of pH
(Plaschke et al., 1999; Balnois et al., 1999), also indicate
an increased surface aggregation with a decrease in pH.
The iron-coagulated sediment should then include some
loosely attached humic colloids that may become rapidly
ARTICLE IN PRESS
Fig. 5. Evolution of elemental ratios as a function of coagulant concentration: N/C (), Fe/C
*
1/10 (&), Na/C
*
1/10 for NHS pH 8
(K) and Ca/C (J).
A.-V. Jung et al. / Water Research 39 (2005) 38493862 3857
volatilized under the electron beam; the microanalysis
being performed mainly with the iron coagulated humic
substances. The almost constant Fe/C atomic ratio
obtained at pH 6 is then in accordance with previous
studies suggesting a complexation mechanism between
oligomeric coagulant species and functional groups of
fulvic and humic colloids around pH 5 (van Breemen et
al., 1979; Jeckel, 1986; Lefebvre and Legube, 1990).
The effect of pH on Fe/C elemental ratio can be
explained by a dynamic reorganization of humic
macromolecules upon interaction with Fe coagulant
species. At pH 8, the behavior of deprotonated humic
material is generally compared with that of a exible
anionic polymer in a stretched conguration (Gosh and
Schnitzer, 1980; Vermeer et al., 1998). In that case, at
low coagulant concentration, the partial neutralization
of humic macromocules by Fe-hydrolyzed species can
induce a reconformation of polymer chains, thus leading
to a compact structure and values of Fe/C close to 3.
Recent uorescence spectroscopy using pyrene as a
polarity-sensitive uoroprobe also suggests such a
rearrangement of humic macromolecules upon associa-
tion with multivalent cations (Engebretson and von
Wandruszka, 1994; Kazpard et al., 2005). At higher
coagulant dosage, the amount of hydrolyzed Fe species
increases relative to the number of humic colloids, and
aggregation between stretched polymer chains may
proceed before folding and reconformation of humic
material take place. In that case, the coagulant species
may form bridges between stretched humic chains. In
accordance with such an aggregation mechanism, the
increase in sediment volume observed with pH above
OCC, is proportional to C/Fe ratios for both SHS and
NHS. Therefore, the increase in sludge volume at pH 8
is here attributed to a change in shape factor of humic
colloids, and not to the formation of abundant iron
hydroxide precipitate as predicted by a sweep-occula-
tion mechanism (Gregory and Dupont, 2001).
3.5. Coagulation of Moselle river raw water
Experiments with Moselle river water were run in
duplicate, similar trends being observed for the two grab
samples. Four kinds of NOM compounds were identi-
ed from TEM examination (Fig. 6): humic-like colloids
(Fig. 6a (A)), by far the most frequent compounds in the
freeze-dried coagulated sediment, were about 20 nm in
diameter with a N/C ratio in the 0.020.05 range;
proteinic substances (Fig. 6a (B)), appeared as large
grayish areas 300 nm in size with N/C values greater
than 0.1; polysaccharidic material was characterized by
micron size spheroids with neither nitrogen nor iron
detected, a CEEL spectrum equivalent to that of a sugar
backbone, and it was easily degraded under the electron
beam (Fig. 6b); small organic colloids intimately
associated with smectite clay material and with a sharp
299 eV peak, were also found (Fig. 6c). Such organic
matter is reminiscent of that observed by Furukawa in
Jourdan river sediments (Furukawa, 2000). It should be
noted that, for NOM, the interpretation of EELS
spectra at the carbon edge is complicated by the
presence of L23 peaks at 297 and 300 eV of potassium
(Henke et al., 1993).
Fig. 7 shows the jar-test results obtained with Moselle
river water coagulated with iron at natural pH (7.4). The
supernatant turbidity continuously increases with iron
concentration up to [Fe]1.2 10
4
mol/L, then steeply
decreases at OCC ([Fe] 1.7 10
4
mol/L), slightly re-
increases above OCC, and then decreases again at higher
coagulant dosage. The evolution with coagulant con-
centration of C-EELS spectra corresponding to humic-
like colloids is illustrated in Fig. 8a. This kind of
material shows a similar reactivity to that of SHS and
NHS at pH 8, with a strong decrease of the p*CQO
peak at 288 eV with iron concentration. Fig. 8b presents
the variation of N/C, Ca/C, and Fe/C elemental ratios of
coagulated NOM with iron concentration. Fe/C in-
creases continuously with coagulant dosage to reach a
value close to 1 above OCC, which is consistent with the
results obtained with SHS and NHS. Likewise, Ca/C
elemental ratio decreases in accordance with a substitu-
tion of Ca for Fe-hydrolyzed species onto functional
groups of organic matter. N/C values reveal that NOM
removal occurs in two main stages: up to OCC, the N/C
atomic ratios are less than 0.03, thus suggesting a
preferential coagulation of humic substances; above
OCC, N/C increases signicantly to reach values in the
0.1 range which indicates the presence of proteinic
matter in the coagulated sediment. Such results are
consistent with previous pyrolysis-gas chromatography-
mass spectrometry (Py-GC-MS) experiments showing
that polyhydroxyaromatics are preferentially removed
by ferric chloride (Vilge Ritter et al., 1999a, b; Jung,
2004). More generally, the hydrophobic fraction of
NOM is known to be more easily destabilized than
hydrophilic compounds (e.g., proteinic matter) (Croue
et al., 1993; Bose and Reckhow, 1998). In particular, the
slight increase of residual turbidity observed above OCC
could correspond to the restabilization of some humic
colloids (Fig. 7).
4. Conclusion
This study shows that TEM observations combined
with EDXS and EELS analyses, represent a powerful
tool to investigate the coagulation of NOM. The
nanoscale resolution of EELS and EDXS is well-
adapted to the size range of organic colloids encoun-
tered in surface waters. In particular, a ngerprinting of
coagulated NOM can be achieved, even though further
ARTICLE IN PRESS
A.-V. Jung et al. / Water Research 39 (2005) 38493862 3858
ARTICLE IN PRESS
250 300 350 400 450
Energy Loss (eV)
250 300 350 400 450
Energy Loss (eV)
285
B
300
346.4
401.6
B : protein-like
substances
B
A
0.5 m
TEM photo
401.6
285
A
289
300
C.K EELS
A : humio-like
substances
A
0
1
2
6 7
8
9
10
0 : C K
a
1' : Fe L
a
0.7 2.1 3.5 4.9 6.3
Energy (keV)
EDXS
Elmt At%
1
2
6
7
8
9
10
11
O
Na
P
S
Cl
K
Ca
Fe
55.6
3.5
1.0
1.1
2.5
2.6
2.1
31.6
1
2
6
7
8
9
10
11
O
Na
P
S
Cl
K
Ca
Fe
81.6
3.4
1.1
1.1
2.6
2.5
6.1
1.6
Elmt At%
B
0
1'
2
6 7
8 9
10
0 : C K
a
1' : Fe L
a
11
1
0.7 2.1 3.5 4.9 6.3
250 300 350 400 450
285
300
C-K EELS TEM photo
0.5 m
EDXS
0
1
Elmt (K) At%
0
1
C
0
55
45
(indicative)
0.7 2.1 3.5 4.9 6.3
C-K EELS
298
299
285
346.4
300 350 400 450
0.2 m
TEM photo
EDXS
1
0
1'
2
3
4
5
67
8
9
10
11
0.7 2.1 3.5 4.9 6.3
0 : C K
a
1' : O L
a
1
2
6
7
8
9
10
11
O
Na
P
3 Mg
4 Al
5 Si
S
Cl
K
Ca
Fe
58.9
3.4
3.4
6.6
9.5
0.5
0.5
2.8
1.8
2.2
10.5
Elmt At%
Energy (keV)
Energy (keV)
Energy Loss (eV)
Energy Loss (eV)
Energy (keV)
250
(a)
(b)
(c)
Fig. 6. Typical NOM compounds identied by combining TEM observation with EDXS and EELS spectra: (a) Humic-like substances
(A) and protein-like material (B), (b) Polysaccharide-like substances, (c) Organo-mineral species.
A.-V. Jung et al. / Water Research 39 (2005) 38493862 3859
work is still required to improve the distinction of
natural organic substances.
EELS and EDXS examination of coagulated sediment
also provides a useful insight into the destabilization
mechanism of humic substances and NOM: Fe-hydro-
lyzed species associate with carboxylic groups of humic
colloids at both pH 6 and 8, thus inducing a release of
previously complexed calcium ions. An overall charge
neutralization/complexation mechanism seems to be
responsible for the removal of humic substances in our
experimental conditions as no Fe-hydroxyde precipitate
could be detected in the coagulated sediment. The
association of non-coagulated humic colloids within
ocs may also occur at pH 6. On the other hand, at
alkaline and neutral pH, our results suggest that a
reorganization of humic macromolecules from a
stretched to a coiled conformation takes place upon
interaction with coagulant species. Such phenomenon is
worth exploring, as the optimal coagulant dosage should
then depend on the complex interplay between the rate
of collisions of organic colloids/coagulant species and
the rate of macromolecules reconformation.
Acknowledgments
This work was achieved within the framework of
Zone Atelier du bassin de la Moselle, with nancial
support from Region Lorraine and CNRS (Contrat de
Plan Etat-Re gion and Programme Environnement Vie et
Socie te ). Agence de lEau Rhin-Meuse is gratefully
acknowledged for providing the Field-centrifuge. The
authors would also like to thank T. Jacquin (Agence de
lEau Rhin Meuse, Rozerieules, France) for providing
technical assistance during eld measurements.
References
Andreux, F., Golebiowska, D., Metche, M., 1980. Oxydative
polymerization of O-diphenols in presence or absence of
amino-acids. Topics on (catechol-glycocolle) and (catechol-
diglycylglycine). General Assembly of Polyphenols Group
Report, Logrono, Spain, Bulletin, vol. 9, pp.178188.
Balnois, E., Wilkinson, K.J., Lead, J.R., Bufe, J., 1999.
Atomic force microscopy of humic substances: effects of pH
and ionic strength. Environ. Sci. Technol. 33 (21),
39113917.
Bose, P., Reckhow, D.A., 1998. Adsorption of natural organic
matter on preformed aluminium hydroxide ocs. J. Envir-
on. Eng. 124, 803810.
Collins, M.R., Amy, G.L., King, P.H., 1985. Removal of
organic matter in water treatment. J. Environ. Eng. 11,
850854.
ARTICLE IN PRESS
N
o
r
m
a
l
i
z
e
d
i
n
t
e
n
s
i
t
y

(
A
.
U
.
)
NOM natural pH 7.4
0.08
0.1
0.12
(v)
(r)
(s)
(t)
(u)
Energy Loss (eV)
250 350 400 450
(a)
0
0.02
0.04
0.06
1.1 10
-4
2.3 10
-4
1.4 10
-4
1.7 10
-4
2.0 10
-4
Iron concentration (M)
E
l
e
m
e
n
t
a
l

a
t
o
m
i
c

t
a
l

(r)
(s)
(t)
(u)
(v)
(b)
300
NOM natural pH 7.4
Fig. 8. (a) EEL spectra for NOM as a function of coagulant
dosage. (b) Elemental atomic ratio: N/C (), Fe/C
*
1/10 (&),
Na/C (K) and Ca/C (J).
0
1
2
3
4
5
6
7
8
9
10
0
10
-4
1.5 10
-4
2 10
-4
2.5 10
-4
NOM natural pH 7.4
(r)
(s)
(t)
(u)
(v)
Iron concentration (M)
5 10
-5
Fig. 7. Jar-test results for NOM at natural pH (7.4). (K
residual turbidity (NTU) sediment volume (mL)).
A.-V. Jung et al. / Water Research 39 (2005) 38493862 3860
Collins, M.R., Amy, G.L., Steelink, C., 1986. Molecular weight
distribution, carboxylic acidity and humic substances
content of aquatic organic matter: implications for removal
during water treatment. Environ. Sci. Technol. 20 (10),
10281032.
Croue , J.-P., Lefebvre, E., Martin, B., Legube, B., 1993.
Removal of dissolved hydrophobic and hydrophilic organic
substances during coagulation/occulation of surface
waters. Water Sci. Technol. 27, 143152.
Dempsey, B.A., Ganho, R.N., OMelia, C.R., 1984. The
coagulation of humic substances by means of aluminium
salts. J. Am. Water Works Assoc. 74, 141150.
Edwards, G.A., Amirtharajah, A., 1985. Removing color
caused by humic acids. J. Am. Water Works Assoc. 77,
5057.
Egerton, R.F., 1996. Electron Energy-Loss Spectroscopy in the
Electron Microscope, second ed. Plenum Press, New York
423pp.
El Samrani, A.G., Lartiges, B.S., Montarge` s-Pelletier, E.,
Kazpard, V., Barre` s, O., Ghanbaja, J., 2004. Clarication
of municipal sewage with ferric chloride: the nature of
coagulant species. Water Res. 38 (3), 756768.
Engebretson, R.R., von Wandruszka, R., 1994. Microorganiza-
tion in dissolved humic acids. Environ. Sci. Technol. 28 (11),
19341941.
Furukawa, Y., 2000. Energy-ltering transmission electron
microscopy (EFTEM) and electron energy-loss spectro-
scopy (EELS) investigation of clay-organic matter aggre-
gates in aquatic sediments. Org. Geochem. 31 (7/8),
735744.
Gosh, K., Schnitzer, M., 1980. Macromolecular structures of
humic substances. Soil Sci. 129, 266276.
Gregory, J., Dupont, V., 2001. Properties of ocs produced
by water treatment coagulants. Water Sci. Technol. 37,
3543.
Gregor, J.E., Nokes, C.J., Fenton, E., 1997. Optimizing natural
organic matter removal from low turbidity waters by
controlled pH adjustment of aluminum coagulation. Water
Res. 31 (12), 29492958.
Gu, B., Schmitt, J., Chen, Z., Liang, L., McCarthy, J.F., 1994.
Adsorption and desorption of natural organic matter on
iron oxide: mechanisms and models. Environ. Sci. Technol.
28 (1), 3846.
Henke, B.L., Gullikson, E.M., Davis, J.C., 1993. X-ray
interactions: photoabsorption, scattering, transmission,
and reection at E 5030000 eV, Z 192. At. Data
Nucl. Data Tables 54, 181342.
Hung, W.T., Chang, I.L., Lin, W.W., Lee, D.J., 1996.
Unidirectional freezing of waste-activated sludges: effects
of freezing speed. Environ. Sci. Technol. 30 (7), 23912396.
Jeckel, M.R., 1986. Interactions of humic acids and aluminium
salts in the occulation process. Water Res. 20 (12),
15351542.
Jung, A.-V., 2004. Interactions des micropolluants anthropi-
ques avec la matie` re organique naturelle et devenir lors des
ope rations de traitement de leau. Ph.D. Thesis INPL,
Nancy, 427pp.
Jung, A.-V., Chanudet, V., Ghanbaja, J., Lartiges, B.S.,
Abdelmoula, M., Bersillon, J.-L., 2005a. Iron populations
evolution during organic matter coagulation by a nitrate
salt: energy-loss spectroscopy (EELS) and Mo ssbauer
investigations. C.R. Acad. Sci. Chim. submitted for
publication.
Jung, A.-V., Frochot, C., Parant, S., Lartiges, B., Selve, C.,
Viriot, M.-L., Bersillon, J.-L., 2005b. Synthesis of amino-
phenolic humic-like substances and comparison with
natural aquatic humic acids: a multi-analytical techniques
approach. Org. Geochem. in press.
Kazpard, V., Lartiges, B.S., Jung, A-V., Frochot, C., Espinose
de la Caillerie, J.B., Portal, J.M., Go rner, T., Viriot, M.L.,
Bersillon, J.L., 2005. Aggregation of a humic macromole-
cules with polycations (Al
13
and PEI): a combined pyrene
uorescence, electrophoretic mobility and SEC investiga-
tions. Water Res. submitted for publication.
Lartiges, B.S., Bottero, J.Y., Derrenger, L.S., Humbert, B.,
Tekely, P., Suty, H., 1997. Flocculation of colloidal silica
with hydrolysed aluminium: an
27
Al solid state NMR
investigation. Langmuir 13 (2), 147152.
Lead, J.R., Wilkinson, K.J., Starchev, K., Canonica, S., Bufe,
J., 2000. Determination of diffusion coefcients of humic
substances by uorescence correlation spectroscopy: role of
solution conditions. Environ. Sci. Technol. 34 (7),
13651369.
Leenheer, J.A., Croue , J.-P., 2003. Characterizing aquatic
dissolved organic matter. Environ. Sci. Technol. 1 (1),
19A26A.
Lefebvre, E., Legube, B., 1990. Iron (III) coagulation of humic
substances extracted from surface waters: effect of pH and
humic substances concentration. Water Res. 24 (5),
591606.
Matilainen, A., Lindqvist, N., Korhonen, S., Tuhkanen, T.,
2002. Removal of NOM in the different stages of the water
treatment process. Environ. Int. 28 (11), 457465.
Maurice, P.A., Namjesnik-Dejanovic, K., 1999. Aggregates
structures of sorbed humic substances observed in aqueous
solution. Environ. Sci. Technol. 33 (9), 15381541.
Myneni, S.C.B., 2002. Soft X-ray spectroscopy and spectro-
microscopy studies of organic molecules in the environ-
ment. Reviews in Mineralogy and Geochemistry
(Applications of Synchrotron Radiation in Low-Tempera-
ture Geochemistry and Environmental Sciences) 49,
485579.
Oliver, B.G., Visser, S.A., 1980. Chloroform production from
the chlorination of aquatic humic material: the effect of
molecular weight, environment and season. Water Res. 14
(9), 11371141.
Owen, D.M., Amy, G.L., Chowdhury, Z.K., Paode, R.,
McCoy, G., Viscosil, K., 1995. NOM characterization and
treatability. J. Am. Water Works Assoc. 87, 4663.
Plaschke, M., Ro mer, J., Klenze, R., Kim, J.I., 1999. In situ
AFM study of sorbed humic acid colloids at different pH.
Colloids Surfaces A: Physicochem. Eng. Aspects 160,
269279.
Randtke, S.J., 1993. Coagulation of NOM: an overview
of the science and U.S. practice. NOM in drinking
water, origin, characterization and removal. Workshop
Proceedings of AWWA about Natural Organic Matter in
Drinking Water. AWWA Research Fundation Ed.,
pp.155163.
Specht, C., Kumke, M., Frimmel, F., 2000. Characterization of
NOM adsorption to clay minerals by size exclusion
chromatography. Water Res. 34 (16), 40634069.
ARTICLE IN PRESS
A.-V. Jung et al. / Water Research 39 (2005) 38493862 3861
Stevenson, I.L., Schnitzer, M., 1982. Transmission electron
microscopy of extracted fulvic and humic acids. Soil Sci. 133
(3), 179185.
Thieme, J., Schmidt, C., Abbt-Braun, G., Specht, C., Frimmel,
F.H., 2002. X-ray microscopy studies of refractory organic
substances. In: Frimmel, Abbt-Braun, Heumann, Hock,
Lu demann, Spiteller (Eds.), Refractory Organic Substances
in the Environment. Wiley-VCH, Weinheim, pp. 239248.
Thurman, E.M., Malcolm, R.L., 1981. Preparative isolation of
aquatic humic substances. Environ. Sci. Technol. 15 (4),
463466.
Tipping, E., 1993. Modeling the competition between alkaline
earth cations and trace metal species for binding by humic
substances. Environ. Sci. Technol. 27, 520529.
Van Breemen, A.N., Nieuwstad, T.J., Van der MeentOlieman,
1979. The fate of fulvic acids during water treatment. Water
Res. 13 (8), 771780.
Van den Hoop, M.A.G.T., van Leeuwen, H.P., Pinheiro, J.P.,
Mota, A.M., Simo es Gonc- alves, M.L., 1995. Voltametric
analysis of the competition between calcium and heavy
metals for complexation by humic material. Colloids
Surfaces A: Physicochem. Eng. Aspects 95, 305313.
Varlot, K., Martin, J.M., Quett, C., 1998. Physical and
chemical changes in polystyrene during electron irradiation
using EELS in the TEM: contribution of the dielectric
function. J. Microsc. 191 (2), 187194.
Vermeer, A.W.P., van Riemsdijk, W.H., Koopal, L.K., 1998.
Adsorption of humic acid to mineral particles. 1-Specic
and electrostatic interactions. Langmuir 14 (10), 28102819.
Vilge -Ritter, A., Masion, A., Boulange , T., Rybacki, D.,
Bottero, J.Y., 1999a. Removal of natural organic matter
by coagulation-occulation: a pyrolysis-GC-MS study.
Environ. Sci. Technol. 33 (17), 30273032.
Vilge -Ritter, A., Rose, J., Masion, A., Bottero, J.-Y., Laine , J.-
M., 1999b. Chemistry and structure of aggregates formed
with Fe-salts and natural organic matters. Colloids Surfaces
A: Physicochem. Eng. Aspects 147 (3), 297308.
Villemin, G., Mansot, J.L., Watteau, F., Ghanbaja, J., Toutain,
F., 1995. Etude de la biode gradation de la matie` re
organique ve ge tale du sol par la spectroscopie des pertes
de nergie de lectrons transmis (EELS: Electron Energy Loss
Spectroscopy): re partition du carbone, de lazote et e valua-
tion du rapport C-to-N au niveau ultrastructural in situ.
C.R. Acad. Sci. Paris 321 (IIa), 861868.
Watteau, F., Villemin, G., Mansot, J.L., Ghanbaja, J., Toutain,
F., 1996. Localization and characterization by electron
energy loss spectroscopy (EELS) of the brown cellular
substances of beech roots. Soil Biol. Biochem. 28 (10/11),
13271332.
Wilkinson, K.J., Balnois, E., Leppard, G.G., Bufe, J., 1999.
Characteristic features of the major components of fresh-
water colloidal organic matter revealed by transmission
electron and atomic force microscopy. Colloids Surfaces A:
Physicochem. Eng. Aspects 155, 287310.
Further reading
Bruchet, A., Rousseau, C., Mallevialle, J., 1990. Pyrolysis-GC-
MS for investigating high-molecular-weight THM precur-
sors and other refractory organics. J. Am. Water Works
Assoc. 71, 6674.
Hitchcock, A.P., Urquhardt, S.G., Rightor, E.G., 1992. Inner-
shell spectroscopy of benzaldehyde, terephtalaldehyde,
ethylbenzoate, terephtaloyd chlorides and phosphogen:
models for core excitation of poly(ethylene) terephtalate.
J. Phys. Chem. 96 (22), 87368750.
Johnson, P.N., Amirtharajah, A., 1983. Ferric chloride and
alum as single and dual coagulants. J. Am. Water Works
Assoc. 54, 232239.
Kikuna, J., Tonner, B.P., 1996. XANES spectra of a variety of
widely used organic polymers at the C K-edge. J. Electron
Spectrosc. Relat. Phenom. 82 (1/2), 5360.
Lefebvre, E., Legube, B., 1993. Coagulation-occulation by
ferric chloride of some organic compounds in aqueous
solutions. Water Res. 27 (3), 433447.
OMelia, C.R., 1972. Coagulation and occulation. In: Weber,
Jr., W. (Ed.), Physicochemical Processes for Water Quality
Control. Wiley Interscience, New York 14pp.
Packam, R.F., 1965. Some studies of the coagulation of
dispersed clays with hydrolysing salts. J. Colloid Sci. 20,
8192.
Rose, J., Vilge , A., Lauquet-Olivie, G., Masion, A., Frechou,
C., Bottero, J.Y., 1998. Iron speciation in natural organic
matter colloids. Colloids Surfaces A: Physicochem. Eng.
Aspect 136 (1/2), 1119.
Semmens, M.L., Field, T.K., 1980. Coagulation: experiences in
organics removal. J. Am. Water Works Assoc. 72, 476483.
Stumm, W., OMelia, C.R., 1968. Stoechiometry of coagula-
tion. J. Am. Water Works Assoc. 60, 514539.
ARTICLE IN PRESS
A.-V. Jung et al. / Water Research 39 (2005) 38493862 3862