Carbamazepine in water: persistence in the environment, ozonation treatment and preliminary assessment on algal toxicity.

Article abstract of DOI:10.1016/S0043-1354(01)00500-0

The presence of carbamazepine (CBZ), an antiepilectic drug, has been reported in sewage treatment plant (STP) effluents as a result of its low biodegradability. In the present work, the persistence of CBZ in aquatic environment with respect to abiotic transformation processes along with its toxicity and capability of accumulating in single aquatic organisms (algae) are evaluated. The possibility of removing CBZ from STP effluents is studied by characterizing its ozonation process through the assessment of kinetics and the distribution of oxidation products.

Full text of DOI:10.1016/S0043-1354(01)00500-0

Water Research 36 (2002) 2869–2877
Carbamazepine in water: persistence in the environment,
ozonation treatment andpreliminary assessment on
algal toxicity
Roberto Andreozzia,*, Raffaele Marottaa, Gabriele Pintob, Antonino Polliob
a Dipartimento di Ingegneria Chimica, Facolta di Ingegneria Chimica, Universita degli Studi di Napoli ‘‘Federico II’’,
P. le V. Tecchio, 80-80125 Napoli, Italy
b Dipartimento di Biologia Vegetale, Facolta di Scienze, Universita. degli Studi di Napoli ‘‘Federico II’’, via Foria,
223-80139 Napoli, Italy
Received10 January 2001; accepted19 October 2001
Abstract
The presence of carbamazepine (CBZ), an antiepilectic drug, has been reported in sewage treatment plant (STP)
effluents as a result of its low biodegradability. In the present work, the persistence of CBZ in aquatic environment with
respect to abiotic transformation processes along with its toxicity andcapability of accumulating in single aquatic
organisms (algae) are evaluated. The possibility of removing CBZ from STP effluents is studied by characterizing its
ozonation process through the assessment of kinetics and the distribution of oxidation products. r 2002 Elsevier
Science Ltd. All rights reserved.
Keywords: CBZ; Photodegradation; Ozonation; Drugs; Surface water; Algal bioassay
1. Introduction
tions in STP effluents andtherefore in surface waters.
Whether the presence of these compounds in the
environment may be harmful for man andliving
organisms is still under debate. In fact, although these
substances have the capability of causing biological
effects, no indications are found in the literature about
their persistence in the environment after the dicharge
andthe real interactions they have with living organisms
at the concentrations at which they are foundin surface
waters. With the reference to the persistence, it has been
reportedthat some pollutants may undergo photoche-
mical transformations, via direct light absorption or
sensitizedreactions [8,9] through which their character-
istics can be modified. Only scant indications are found
on their removal from wastewaters [10].
The presence of pharmaceutical compounds in river
andsurface waters has been frequently reportedin the
last years [1–5]. Veterinary andhuman use are generally
recognizedas the most relevant sources for this new
class of environmental pollutants [6]. In fact many of the
numerous active substances usedto treat illnesses are
excretedunmetabolisedor as active metabolites. More-
over, improper disposal of expired medications can
contribute to the contamination of the environment.
Particularly, it has been reportedin Germany the
presence of up to 32 drugs in the influents to Municipal
Sewage Treatment Plants [7]. Some of these compounds
are only partly removedin conventional biological
treatments since they are foundat significant concentra-
Carbamazepine (CBZ) is an antiepileptic drug whose
occurrence in municipal STP effluents andin domestic
wastewaters has been reportedin the literature [7,11].
No indications are available in specialized journals for
its persistence in the environment and/or for its toxicity
*Corresponding author. Tel.: +39-081-768-2251; fax: +39-
81-593-6936.
E-mail address: andreo@irc.na.cnr.it (R. Andreozzi).
0043-1354/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 4 3 - 1 3 5 4 ( 0 1 ) 0 0 5 0 0 - 0

2870
R. Andreozzi et al. / Water Research 36 (2002) 2869–2877
towards simple living organisms such as algae nor for its
removal from aqueous streams by means of alternative
chemical treatments. Recently, some papers have been
reportedon the assessment of the risk of major
congenital abnormalities associatedwith CBZ assump-
tion [12,13].
solution at 36 dm3 hꢀ1
.
For batch experiments
3.0 ꢁ 10ꢀ3 dm3 of ozonatedaqueous solution have
been rapidly mixed with 1.0 ꢁ 10ꢀ1 dm3 of CBZ
(3.3 ꢁ 10ꢀ6 mol dmꢀ3) aqueous solution andthe result-
ing solution kept at rest for 10 min.
CBZ and p-nitroanisole concentrations were deter-
minedby HPLC (1100 Hewlett Packard) using Pheno-
mex C6 reverse phase column with a mixture of water/
acetonitrile (60/40) as mobile phase at 5.0 ꢁ
10ꢀ4 dm3 minꢀ1 andmeasuring the absorbance at
210 nm for CBZ and320 nm for p-nitroanisole.
The pH of aqueous solutions, adjusted with dilute
HClO4 or NaOH, was determined using Orion 960 pH-
meter with a glass pH electrode.
In the present paper, all these aspects are facedaiming
at assessing the persistence of CBZ in the environment
andpreliminarily evaluating its toxicity towards algas.
The capability of ozonation treatment of removing this
species from aqueous streams is also studied.
2. Experimental
The absorbances of humic acid(5.0 ꢁ 10ꢀ3 g d mꢀ3
)
andCBZ (8.0 ꢁ 10ꢀ6 mol dmꢀ3) were measuredwith
2.1. Solar photodegradation and oxidation experiments
UV-VIS spectrophotometer (HP 8452 A) with quartz
cell path length of 1.0 cm.
Solar photodegradation experiments were carried out
during spring at Naples in south of Italy (401N–141E).
They were conducted in 0.5 dm3 annular magnetically
stirredglass reactors refrigeratedat 25 1C [14]. Aqueous
solutions to be irradiated were prepared by dissolving a
proper amount of CBZ in bidistillated water or in
natural waters from Liri river (Frosinone, Italy:
All the reagents but hydrogen peroxide (30% by
weight, Fluka) were purchasedfrom Sigma Aldrich.
For the identification of intermediates and products
of CBZ ozonation the samples withdrawn from the
reactor have been submitted to oxidative, hydrolytic and
derivatizing work-up.
pH=7.8,
Clꢀ=5.1 ꢁ 10ꢀ4 mol dmꢀ3
,
NOꢀ3 =1.0 ꢁ
10ꢀ4 mol dmꢀ3, HCO3ꢀ=5.1 ꢁ 10ꢀ3 mol dmꢀ3).
Solar actinometry was daily carried out by measuring
the concentration decay of a solution of p-nitroanisole
(2.0 ꢁ 10ꢀ6 mol dmꢀ3) in pyridine [15]. The pyridine
concentration was adjusted to obtain during the
irradiation time an almost complete consumption of
p-nitroanisole.
2.2. Oxidative work-up
A drop of concentrated H2O2 solution (10 mol dmꢀ3
)
has been added to an aliquot of ozonated sample
after having adjusted the pH to 10. Resulting
solutions have been heatedfor 2 h at 50 1C. This
treatment allows an easy conversion of aldehydic
function to carboxylic.
According to Dullin and Mill the photolytic degrada-
tion of p-nitroanisole in a batch reactor is described by
means of the following equation:
2.3. Hydrolytic work-up
d½CŠ
dt
ꢀ
¼ qa½CŠ ¼ kp½CŠ
ð1Þ
An ozonated sample, after addition of one drop of
concentratedperchloric acid, has been heatedat 150 1C
in oil-bath for 40 h at reflux. After cooling, the sample is
divided into two parts. The first aliquot has been
submittedto an oxiadtive treatment with H 2O2 for
determining anthranilic acid, the second has been
employedfor the ammonia test.
with the dimensionless parameter q given by
q ¼ 0:44½pyridineŠ þ 0:00028
ð2Þ
and a being dependent on system characteristics (optical
path, volume), sunlight intensity andoptical absorption
properties of the substrate.
Therefore from experimental values of kp the sunlight
intensity is indirectly evaluated by:
2.4. Derivatizing work-up
kp
a ¼
:
ð3Þ
For the analysis of a-dicarbonylic compounds,
1.0 ꢁ 10ꢀ3 dm3 of the ozonatedmixture has been mixed
in a cap-screw vial with 2.0 ꢁ 10ꢀ3 dm3 of a solution
preparedby means of 2.37 ꢁ 10ꢀ1 g of 1,2-phenylene-
diamine and 2.0 ꢁ 10ꢀ3 dm3 of concentratedsulphuric
acidandtaking the volume at 2.5 ꢁ 10ꢀ1 dm3. Samples
have been heatedfor 1 h at 100 1C andafter neutraliza-
tion analysedby HPLC using a mobile 90/10 buffered
solution (H2O:CH3OH:H3PO4, 500:25:2)/CH3CN.
q
In fact, for establishedapparatus andchemical system,
the values assumedby the ratio kp=q depend only upon
the sunlight intensity.
For ozonation experiments with continuous ozone
feeding, a 0.8 dm3 glass-stirredreactor has been em-
ployed[16]. An ozonizedoxygen stream with an ozone
content of 2.0% per volume has been bubbledin the

R. Andreozzi et al. / Water Research 36 (2002) 2869–2877
2871
Evolved carbon dioxide has been trapped by means of
Ba(OH)2 solutions andquantifiedby titration with
standardized HCl solution.
using Student-Newman-Keuel test (SNK), at a ¼ 0:05:
The SPSS statistical package (SPSS Inc. for Windows)
was used.
Oxalic and oxamic acids were detected by HPLC
analyses with an Alltech OA-1000 acidcolumn (mobile
phase: H2O/H2SO4, 500/0.8).
The in vitro long-term effects of CBZ on C 202-7a
have been assessedby exposing the algae to the
compoundfor 60 days. The bioassays have been carried
out in large conical flasks with 1 l of BBM. CBZ
8.0 ꢁ 10ꢀ5 mol dmꢀ3 has been aseptically added and
gently shaken for 3 days before the algal inoculum
(1 ꢁ 105 cells mlꢀ1). The cultures have been incubated
under the above described conditions of light and
temperature, andsamples (2.0 ꢁ 10ꢀ2 dm3) from the
culture batch have been taken at intervals of 3 days. The
samples have been centrifuged(5000 rpm ꢁ 15 m) and
washedtwice with coldBBM (20 ml each). The algal
pellet has been lyophilizedandextractedwith methanol.
CBZ has been measuredin the sample meidum, in
washedsamples andin the algal pellet.
An iodometric procedure [17] has been adopted for
H2O2 analysis.
Colorimetric procedure has been used for the analysis
of ammonia [18]. Nitrate has been reduced to ammonia
by means of Devarda’s alloy in alkaline medium after
heating at 501C for 3 h [19].
2.5. Bioassays
The algae testedagainst CBZ were Ankistrodesmus
braunii, strain CCAP 202-7a, from the Culture Collec-
tion of Algae andProtozoa, Ambleside, Cumbria (UK)
and Selenastrum capricornutum, strain UTEX 1648,
from the Culture Collection of algae at the University
of Texas at Austin (USA). These algae, among the most
commonly usedin laboratory testing, have been grown
in Bold’s basal medium (BBM) [20] in 1.0 dm3 Erlen-
meyer flasks at 23711C with a total irradiance of
100 mE sꢀ1 m2 provided by a daylight fluorescent Philips
lamps (TLD 30w/55). The photoperiodwas 16 h light–
8 h dark. The flasks have been placed on a plexiglas
shaking apparatus as indicated by Shihira and Kraus
[21]. The growth of the alga has been followedat 550 nm
with a colorimeter Bausch & Lomb spectronic 20.
Inocula corresponding to 1 ꢁ 103, 1 ꢁ 104 and1 ꢁ 105
cells mlꢀ1 from axenic cultures in mid-exponential phase
have been grown in test tubes containing BBM and,
respectively, (a) CBZ at concentrations ranging from
2.1 ꢁ 10ꢀ6 g d mꢀ3 (the actual concentration found
in same STP effluents of Germany [7]), to
3. Results and discussion
3.1. Persistence in surface waters
In Fig. 1 the absorption spectrum of CBZ (black
points) is shown which indicates that this molecule
absorbs solar UV radiation. In surface waters it could
thus photolize andundergo photochemical transforma-
tions. This tendency has been checked by means of
proper experiments in which an aqueous solution
containing CBZ at starting concentration of
0.30
0.25
2.0 ꢁ 10ꢀ2 g d mꢀ3
, (b) aqueous solutions of CBZ
(8.0 ꢁ 10ꢀ3 mol dmꢀ3) at different times of ozonation
(0, 5, 10, 20, 30 and40 min). Each test tube containing
the algal inocula anda volume (6.0 ꢁ 10ꢀ3 dm3) of the
testing solution has been incubatedon a shaking
apparatus as previously described. The growth of the
cultures has been followeddaily either by counting the
cell number with a Burker bloodcounting chamber or by
measuring the absorbance increase at 550 nm with a
colorimeter. During the time course of the experiments
(96 h) the algal cultures remainedin the exponential
phase of growth. The tests were carriedout in axenic
conditions. Cells have been thoroughly mixed for 10 s by
a vortex mixer prior to reading. Control containing only
distilled water and BBM was also tested. Growth
experiments have been carriedout five times andresults
have been evaluatedon the basis of the average of five
tests. To test statistical significance of results, one-way
ANOVA has been performedat a ¼ 0:05: For each
solution, a comparison among means was performed
N
0.20
C
O
NH2
0.15
0.10
0.05
0.00
200 250 300 350 400 450 500 550 600
Wavelength (nm)
Fig. 1. Absorption spectra of humic acid(solidline) andCBZ
(black points) in the region 200–600 nm. [humic acid-
s]o=5.0 ꢁ 10ꢀ3 g d mꢀ3; [CBZ]o=8.0 ꢁ 10ꢀ6 mol dmꢀ3
.

2872
R. Andreozzi et al. / Water Research 36 (2002) 2869–2877
1.00
0.80
0.60
0.40
0.20
0.00
1.00
0.80
0.60
0.40
0.20
0.00
2
4
6
8
10
12
14
0
0
20
40
60
80
Irradiating time (hours)
Irradiating time (hours)
Fig. 3. Nitrate-induced photodegradation of CBZ in bi-dis-
tilled water during exposure to sunlight at different concen-
Fig. 2. Photodegradation of CBZ in bi-distilled water during
exposure to sunlight at pH=5.5 and T ¼ 251C.
[CBZ]o=8.0 ꢁ 10ꢀ6 mol dmꢀ3. Start time after 10 a.m.–end
time before 4 p.m. (middle April to early May). Experimental
points (full circle), calculatedvalues with first order kinetic
model (solid line).
trations
of
nitrate
at
pH=5.5
and
T ¼ 251C.
[CBZ]o=8.0 ꢁ 10ꢀ6 mol dmꢀ3. Start time after 10 a.m.–end
time before 4 p.m. (middle April to early May) }, without
nitrate;
E,
nitrate
(5.0 ꢁ 10ꢀ4 g d mꢀ3);
ꢂ,
nitrate
(1.0 ꢁ 10ꢀ2 g d mꢀ3); ’, nitrate (1.5 ꢁ 10–2 g d mꢀ3).
8.0 ꢁ 10ꢀ6 mol dmꢀ3 has been submittedto solar irra-
diation in the apparatus described in the experimental
section. The results of these runs are reportedin Fig. 2
(full circles) andindicate that the substrate concentra-
tion really decreases with increasing the irradiation time.
Actinometric measurements with p-nitroanisole in pyr-
idine performed each day in a similar reactor with the
same exposure with respect to solar irradiation gave an
average value of kp=q ¼ 72:5 hꢀ1: To fully assess the
nature of the observedreaction, an aqueous solution
with the same concentration of CBZ has been kept in the
dark and analyzed at different reaction times. These
results (not shown) indicated that no reaction occurs
without irradiation. The data reported in Fig. 2 for the
photochemical degradation of CBZ have been analyzed
by plotting the natural log of the relative residual
concentration against the time andthe pseuod first-
order kinetic constant derived by means of a least-
squares regression (solidline). A kinetic constant value
of 5.7 ꢁ 10ꢀ3 hꢀ1 has been obtainedfrom which a half-
life t1=2 ¼ 121:6 h has been calculated.
runs have thus been performed by adding to CBZ
aqueous solutions nitrates andhumic acid(separately).
CBZ half-life reduced to 69.0, 24.5 and 11.2 h when
the concentrations of nitrates were, respectively, of
5.0 ꢁ 10ꢀ4 g d mꢀ3
10ꢀ2 g d mꢀ3 (Fig. 3).
The presence
,
1.0 ꢁ 10ꢀ2 g d mꢀ3
and1.5
ꢁ
of dissolved
humic
acid
(5.0 ꢁ 10ꢀ3 g d mꢀ3) on the other hands resulted into an
increase of CBZ half-life (233.7 h) (Fig. 4). The addition
of these compounds to CBZ aqueous solutions seems
to hinder the spontaneous photochemical degradation
of the pollutant.
A possible explanation for this
behaviour couldbe foundby consiedring that the
adopted humic acid is characterized by strong UV
absorptions in the same range as CBZ (Fig. 1). That is,
when humic acidis present, only a part of the irradiated
power is absorbedby CBZ itself which undergoes a
slower photochemical transformation with respect to
that observedin bidistilledwater. A similar behaviour
has been observed by others for the degradation of
pesticides [28,29].
Several authors reportedabout the capability of
humate of generating, after sunlight absorption, reactive
oxidant species, such as singlet oxygen [22,23] and
hydroxyl radicals [24] or of enabling rapid photosensi-
tizedreactions of certain pollutants via energy transfer
from molecules in their triplet states [25]. Nitrite and
nitrate are also indicated as capable of promoting the
formation of OH radicals [26,27]. Some experimental
To fully assess the fate of CBZ in real natural waters,
aqueous solutions preparedby spiking river water with
this substance have been submittedto solar radiation. In
Fig. 5 the relative residual concentration is shown
against the time. From these data an half-life of 907
sunlight hours is evaluatedwith an average kp/q for the
system p-nitroanisole equal to 26.8 hꢀ1
.

R. Andreozzi et al. / Water Research 36 (2002) 2869–2877
2873
whereas in the blank andin all the cultures of S.
capricornutum CBZ concentration remainedconstant.
These results indicate that CBZ exerts only negligible
effects on microalgae within 96 h. However, a toxicity of
the substance over longer times could not be excluded.
Experiments with 60 days CBZ expo-sure were planned
on A. braunii, to determine if this alga was capable of
accumulating the substrate by adopting high concentra-
tions in comparison with the culture medium. The
uptake andconcentration into algal cells of CBZ could
cause secondary sensitivity of alga to this compound or
bioaccumulation through foodnet to consumer organ-
isms in the upper trophic level.
1.00
0.80
0.60
0.40
0.20
0.00
In a 60-day experiments the initial algal concentration
was 1 ꢁ 105 cells mlꢀ1 andCBZ has been axenically added
to give a concentration of 1.9 ꢁ 10ꢀ2 g d mꢀ3. During
these tests growth rate, morphology of cells, number and
shape of sporangia andautospores have been monitored.
In the blank cultures as well as in CBZ-treatedones the
stationary phase of growth was reachedafter 15 days. In
this phase the algae presentednumerous droplets and
grains in the protoplasm, andcells of irregular shape were
frequently observed. Moreover, algal cultures showed an
increasing number of sporangia. A comparison with the
sporangia observedin the exponential cultures evidenced
that in the stationary phase the sporangia were larger and
formedmore autospores. These changes can be attributed
to the senescence of algal population [30] andno
significant differences between control and CBZ-treated
cells were observed
0
20
40
60
80
Irradiating time (hours)
Fig. 4. Humic acids effect on the photodegradation of CBZ
in bi-distilled water during exposure to sunlight at pH=5.5
and T ¼ 251C. [CBZ]o=8.0 ꢁ 10ꢀ6 mol dmꢀ3. Start time after
10 a.m.–end time before 4 p.m. (middle April to early May)
E without humic acids; ꢂ humic acids (5.0 ꢁ 10ꢀ3 g d mꢀ3).
1.00
0.80
0.60
0.40
0.20
0.00
On the other hand, the concentration of CBZ
progressively decreased in the algal culture (Fig. 6),
and, after 60 days, over 50% of CBZ was disappeared in
1.0
0.8
0.6
0.4
0.2
0.0
0
200
400
600
Time (daylight hours)
Fig. 5. Phototransformation of CBZ in river water during
exposure to spring sunlight (23/03/00–31/05/00). [CBZ]o=
8.0 ꢁ 10ꢀ6 mol dmꢀ3ꢀT=251C.
3.2. Algal bioassays
0
10
20
30
Time (dd)
Fig. 6. Transformation of CBZ in Bold’s basal medium (BBM)
40
50
60
70
In the bioassays no inhibition of growth, measuredusing
biomass variation after 96 h from the addition of CBZ, has
been found, independently of algal and CBZ concentra-
tions. At the endof the experiment a slight decrease of CBZ
in the medium has been observed in A. braunii cultures,
without algae (’) andin algal culture
8.0 ꢁ 10ꢀ5 mol dmꢀ3 ꢀT=251C.
( ꢂ). [CBZ]o=

2874
R. Andreozzi et al. / Water Research 36 (2002) 2869–2877
the medium. No significant amounts of CBZ could be
detected in A. braunii cells during the course of the
experiment. As previously reported, the strain 202.7a of
A. braunii is able to subtract from the medium a-
asarone, a toxic phenolic substance, andno significant
amount of this compoundcouldbe detectedin the cell
[31]. In the present case, it couldbe put forwardthat
CBZ is taken up by algal cells andenter into biochemical
processes. The utilization of unusual compounds as
metabolites has been described for other unicellular
algae. It has been reported[32] that the redalga
Galdieria sulphuraria couldbe grown on a wide range
of rare sugar-alcohols, whilst Sample andCain [33]
dosage of 2.08 ꢁ 10ꢀ5 mol dmꢀ3, a ratio of approxi-
mately 10 (ozone moles/CBZ moles) is realized. A
complete abatement of CBZ has been recorded in all the
performedruns.
This result indicates that a complete removal of this
pollutant may be easily achieved in drinking waters, due
to the fact that expectedCBZ concentrations in real
water streams are below the values adopted in these tests
[7].
A kinetic analysis has been also attemptedby using
the results collectedduring semi-continous ozonation
runs. Since negligible ozone concentrations have been
foundin the liquidbulk during the early ozonation
stages, the process has been considered to develop under
a ‘‘quasi-diffusive’’ regime of absorption with reaction.
A proper fluid-dynamic submodel reported elsewhere
[35] has been coupledwith a simplifiedkinetic one in which
a single overall reaction was usedto describedthe ozone
attack to the substrate in the initial ozonation stages:
reportedthat the chrysophyte Ochromonas danica
metabolizedphenol by the metacleavage pathway.
3.3. Oxidative treatments
The results reportedin the previous paragraphs show
that CBZ is characterizedby a certain persistence in the
environment andby the absence of toxic effects towards
andaccumulating capability in testedaquatic organisms
(algae).
kO3
CBZ þ zO3 - products:
The adopted model has been used to estimate the best
values of kO and z (Table 1) by means of a suitable
optimizing procedure [36].
Since these results cannot rule out the existence of some
toxic effects exertedby this species towards other living
organisms [34], a conservative approach to the problem
of the presence of CBZ in the environment wouldthus
suggest its removal at least from water streams for human
use. To this purpose some investigations have been firstly
performedto assess if CBZ couldbe destroyedduring the
oxidative treatments to which drinking waters are
currently submitted[10]. When ozonation is employed,
an ozone dosage of 1.0ꢁ 10ꢀ3 g d mꢀ3 anda residence
time of 10 min are adopted.
3
Some attempts to identify, during the above-reported
ozonation runs, the intermediates and products of CBZ
oxidation failed probably because of the low concentra-
tions at which they were present in the reacting solution.
To more deeply investigate this point a second set of
ozonation experiments have been performedby using
CBZ
solution
at
starting
concentration
of
5.0 ꢁ 10ꢀ4 mol dmꢀ3
.
Due to the electrophilic nature of the ozone attack to
the substrate, two pathways can be foreseen for the
ozonation process depending on the centre to which the
oxidant initially binds:
Some ozonation runs have thus been carriedout,
according to the procedure described in Section 2, by
using CBZ aqueous solutions at starting concentration
of 3.3 ꢁ 10ꢀ6 mol dmꢀ3. In this way, with an ozone
_
O 3
1
+
N
C
O
( I )
O3
NH2
N
_
O 3
C
O
+
NH2
2
N
C
O
NH2
( II )

R. Andreozzi et al. / Water Research 36 (2002) 2869–2877
2875
Table 1
Kinetic rate constant (kO ), stoichiometric coefficient (z) and
Since for species (I) a benzylic-like cation structure
can be written among the others too,
3
percentage standard deviations (% s) obtainedfrom experi-
mental runs in bi-distillated water. T=25.01C, [CBZ]o=5.0 ꢁ
10ꢀ4 mol dmꢀ3
_
O 3
+
(molꢀ1 dm3 sꢀ1
)
7.81 ꢁ 10471.31 ꢁ 104
k
O3
z
9.85 ꢁ 10ꢀ1 72.115 ꢁ 10ꢀ2
4.7
3.7
N
(%)
sCBZ (%)
sO
3
C
O
NH2
sence of anthranilic acid(Fig. 7). Successive analysis on
samples hydrolyzed and derivatized with 1,2 phenylene-
diamine allowed the identification of glyoxal, glyoxilic
andoxalic acid(Fig. 7).
it is more stabilizedby resonance than species (II),
suggesting that the oxidation process proceeds mainly
through path 1 with minor occurrence of path 2.
According to both path 1 and path 2 the formation of
an anthranilic acidprecursor,
From the diagrams of Fig. 7 it is clear that a low
degree of mineralization is achieved for even long
ozonation times, evolved carbon dioxide account-
ing for up to 30% of the initial carbon content.
Moreover, total carbon balance results lacking even
for prolongedozonation thus indicating the presence in
the reacting solution of some non-identified compounds.
Since the formation amidic intermediates can be
predicted in the ozonation process [37], hydrolytic
treatments have been carriedout on ozonatedsamples
to favour the release of ammonia. Ammonia determina-
tion on aqueous samples starting with initial CBZ
concentration equal to 5.0 ꢁ 10ꢀ4 mol dmꢀ3 withdrawn
after 60 and90 min of ozonation account for up to,
respectively, 68.9% and100% of the initial nitrogen
content. No nitrates have been foundudring the
ozonation of CBZ since ammonia measurements gave
the same results after reductive treatment with Devar-
da’s alloy.
O
C
(III)
X
X
N
X
C
O
C
O
OH H
,
O
C
NH2
C1 (CO2), C2 (glyoxal, glyoxilic, oxalic andoxamic
acids) and C3 compounds (ketomalonic acid) can be
predicted.
HPLC analyses performedon directly injectedsam-
ples indicated a complete disappearance of the substrate
after 4 min of ozone feeding with a specific ozone
consumption of 1.0 mol molꢀ1 of CBZ andthe pre-
senceFat trace levelFof oxamic acid. Since no
standard was available for the species (III) it has been
considered that it gives rise to the formation of
anthranilic acid after an hydrolitic and oxidative
treatment:
Test for toxicity assessment of CBZ solutions after
treatment with O3 at different reaction times (0, 5, 10,
20, 30 and40 min) have been carriedout on
S.
capricornutum and A. braunii. The duration of the test
rangedfrom 96 h to 30 days. No inhibition has been
observed.
The analyses performedon samples submittedto
hydrolytic and oxidative work-up confirmed the pre-
O
C
+
X
H
H
2 O
,
X
NH3
HO
( III )
+
+
CO2
+
C
O
C
O
NH2
O
C
O
_
C
_
X
HO
H
2 O
,
O
2
∆
NH2
NH2

2876
R. Andreozzi et al. / Water Research 36 (2002) 2869–2877
References
1.20
1.00
0.80
0.60
0.40
0.20
0.00
100
80
60
40
20
0
[1] Stumpf M, Ternes TA, Haberer K, Seel P, Baumann W.
Determination of pharmaceutics in sewage plants andriver
water. Vom Wasser 1996;86:291–303.
[2] Sacher F, Lochov E, Bethmann D, Brauch HJ. Occurrence
of drugs in surface waters. Vom Wasser 1998;90:233–43.
[3] Ternes TA, Stumpf M, Schuppert B, Haberer K.
Simultaneous determination of antiseptics and acidic
drugs in sewage and river water. Vom Wasser
1998;90:295–309.
[4] Hirsch R, Ternes T, Haberer K, Kratz KL. Occurrence of
antibiotics in the aquatic environment. Sci Total Environ
1999;225:109–18.
[5] Zuccato E, Calamari D, Natangelo M, Fanelli R. Presence
of therapeutic drugs in the environment. Lancet
2000;355:1789–90.
0
10
20
30
40
50
60
Time (min)
[6] Halling-Sorensen B, Nielsen SN, Lanzky PF, Ingerslev F,
Lutzhorft HC, Jorgensen SE. Occurrence, fate andeffects
of pharmaceutical substances in the environmentFa
review. Chemosphere 1998;36:357–93.
Fig. 7. Ozonation of CBZ andreaction intermeidates at
pH=5.5. [CBZ]o=5.0 ꢁ 10ꢀ4 mol dmꢀ3–T ¼ 251C ’ CBZ, ꢂ,
hydrogen peroxide, E, oxalic acid, J, glyoxylic acid, n,
carbon dioxide, +, glyoxal, %, oxamic acid, m, anthranilic
acidand &, ketomalonic acid.
[7] Ternes TA. Occurrence of drugs in German sewage
treatment plants andrivers. Water Res 1998;32:3245–60.
[8] Stangroom SJ, MacleodCL, Lester JN. Photosensitized
transformation of the herbicide 4-chloro-2-methylphenoxy
acetic acid(MCPA) in water. Water Res 1998;32(3):
623–32.
[9] Mateus DA, Da Silva AM, Burrows HD. Kinetics of
photodegradation of the fungicide fenarimol in natural
waters andin various salt solutions: salinity effects
andmechanistic consiedrations. Water Res 2000;34(4):
1119–26.
4. Conclusions
The persistence in the environment of CBZ, its
preliminary toxicity towards algas and its oxidative
degradation by means of ozone have been studied in the
present work. Collectedresults edmonstrate that the
studied substrate is capable of photolyzing in distilled
andriver waters. Nitrate andhumic acidhave opposite
effects on its degradation, the first promoting the second
inhibiting. These preliminary bioassays seems to indicate
that CBZ is not toxic towards Selenastrum capricornu-
tum and Ankistrodesmus braunii. Moreover, no accumu-
lation of the compoundinto the algal cells have
been observed. However, further experiments on other
test organisms belonging to algae andother aquatic
species are requiredto fully assess the toxicity of CBZ in
the environment. The ozonation has been demonstrated
to be a suitable tool for CBZ abatement even at the
process conditions usually adopted in drinking water
facilities. The product distribution during ozonation has
been also investigated. A low degree of mineralization is
observedafter 60 min of ozonation treatment. The
carbon balance remains lacking for even long ozonation
times.
[10] Zwiener C, Frimmel FH. Oxidative treatment of pharma-
ceutical in water. Water Res 2000;34(6):1881–5.
[11] Seiler RL, Zaugg SD, Thomas JM, Howcroft DL. Caffeine
andpharmaceuticals as indicators of waste water contam-
ination in wells. GroundWater 1999;37(3):405–10.
[12] Samren EB, Van Duijn CM, Christiaens GC, Hofman A,
Lindhout D. Antiepileptic drug regimens and major
congenital abnormalities in the offspring. Ann Neurol
1999;46(5):739–46.
[13] Olney JW, Farber NB, Wozniak DF, Jevtovic-Todorovic
V, Ikonomidou C. Environmental agents that have the
potential to trigger massive apoptotic neurodegeneration
in the developing brain. Environ Health Perspectives
2000;108(Suppl. 3):383–8.
[14] Andreozzi R, Caprio V, Insola A, Marotta R. The
oxidation of metol (N-methyl-p-aminophenol) in aqueous
solution by UV/H2O2 photolysis. Water Res
2000;34(2):463–72.
[15] Dullin D, Mill T. Development andevaluation of sunlight
actinometers. Environ Sci Technol 1982;16:815–20.
[16] Caprio V, Insola A. Aniline andanilinium ion ozonation
in aqueous solution. Ozone Sci Eng 1985;7:169–73.
[17] Treadwell FP, Hall J. Analytical chemistry, vol. II, 8th ed.
New York: Wiley, 1935. p. 117.
Acknowledgements
[18] Weatherburn M. Phenol-hydrochlorite reaction for deter-
mination of ammonia. Anal Chem 1967;39(8):971–4.
[19] Standard Methods for the examination of water and
wastewater. 16th ed. American Public Health Association,
Washinghton, DC, 1985. p. 391.
The authors wish to thank the Commission of the
European Communities for the financial support of this
work under Grant No EVK1-CT-2000-00048

R. Andreozzi et al. / Water Research 36 (2002) 2869–2877
2877
[20] Nichols HW. Growth media-freshwater. In: Stein JR,
editor. Handbook of phycological methods. Cambridge:
Cambridge University Press, 1973. p. 7–24.
the presence of humic substances. Fresenius Environ Bull
1999;8(7–8):478.
[30] Legnerova J. The systematic andontogenesis of the
genera Ankistrodesmus Corda and Monoraphidium
gen. nov. In: Fott B, editor. Studies in phycology.
Stuttgart: E. Schweizerbart’sche Verlagsbuchhandlung,
1965. p. 75–144.
[21] Shihira I, Kraus RW. Chlorella: physiology andtaxonomy
of forty-one isolates. Baltimore: University of Maryland
Press, 1965. p. 1–97.
[22] Zepp RG, Wolfe NL, Baughman GL, Hollis RC. Singlet
oxygen in natural waters. Nature 1977;267:421–3.
!
[23] Haag WR, Hoigne J. Singlet oxygen in surface waters. 3.
[31] Pollio A, Pinto G, Ligrone R, Aliotta G. Effects of the
potential allelochemical a-asarone on growth, physiology
andultrastructure of two unicellular green algae. J Appl
Phycol 1993;5:395–403.
Photochemical formation andsteady-state concentrations
in various types of waters. Environ Sci Technol
1986;20:341–8.
[32] Gross W, Schnarrenberger C. Heterotrophic growth of
two strains of the acido-thermophilic red alga Galdieria
sulphuraria. Plant Cell Physio. 1995;36(4):633–8.
[33] Sample KT, Cain RB. Biodegradation of phenols by the
alga Ochromonas danica. Appl Environ Microbiol
1996;62(4):1265–73.
[24] Vaughan PP, Blough NV. Photochemical formation of
hydroxyl radical by constituents of natural waters. Environ
Sci Technol 1998;32:2947–53.
[25] Faust BC. In: Helz GR, Zepp RG, Crosby DG, editors.
Aquatic,surface photochemistry. Ann Arbor MI: Lewis
Publishers, 1994. p. 3–39.
[34] Kenaga EE, Moolenaar RJ. Fish andDaphnia toxicity as
surrogates for aquatic vascular plants andalgae. Environ
Sci Technol 1979;13(12):1479–80.
!
[26] Zepp RG, Hoigne J, Bader H. Nitrate-induced photoox-
idation of trace organic chemicals in water. Environ Sci
Technol 1987;21:443–50.
[35] Andreozzi R, Caprio V, Insola A, Tufano V. Measuring
ozonation rate constants in gas–liquidreaction under the
kinetic-diffusional transition regime. Chem Eng Commun
1996;143:195–7.
[27] Zafiriou O, Bonneau R. Wavelength-dependent quantum
yield of hydroxyl radical formation from photolysis of
nitrite ion in water. Photochem Photobiol 1987;45:723–7.
[28] Taboada RE, Schmitt P, Mansour M. Photodegradation
of pirimicardin natural water andin different aqueous
solutions under simulated sunlight conditions. Fresenius
Environ Bull 1999;4(11):649–52.
[36] Reklaitis GV, Ravindran A, Regsdell KM. Engineering
optimization. New York: Wiley, 1983.
[37] Andreozzi R, Caprio V, D’Amore MG, Insola A, Tufano
V. Analysis of complex reaction networks in gas–liquid
systems: the ozonation of 2-hydroxypyridine in aqueous
solutions. IndEng Chem Res 1991;30:2098–104.
[29] Elazzouzi M, BensaoudA, Bouhaouss A, Guittonneau S,
Dahchour A, Meallier P. Photodegradation of imazapyr in