Water Research 38 (2004) 414–422

ARTICLE IN PRESS

*Correspond

+39081593693

E-mail addr

0043-1354/$ - se

doi:10.1016/j.w

Advanced oxidation of the pharmaceutical drug diclofenacwith UV/H2O2 and ozone

Davide Vognaa, Raffaele Marottab, Alessandra Napolitanoa,Roberto Andreozzib,*, Marco d’Ischiaa

aDip. di Chimica Organica e Biochimica, Federico II, Facolta di Scienze, Universita degli Studi di Napoli Complesso Universitario M. S.

Angelo, via Cinthia, 4, 80126, Napoli, ItalybDip. di Ingegneria Chimica, Facolta di Ingegneria, Universita degli Studi di Napoli FedericoII Prazzale Tecchio, 80, 80125 Napoli, Italy

Received 24 October 2002; received in revised form 2 September 2003; accepted 19 September 2003

Abstract

Diclofenac, a widely used anti-inflammatory drug, has been found in many Sewage Treatment Plant effluents, rivers

and lake waters, and has been reported to exhibit adverse effects on fish. Advanced oxidation processes, ozonation and

H2O2/UV were investigated for its degradation in water. The kinetic of the degradation reaction and the nature of the

intermediate products were still poorly defined. Under the conditions adopted in the present study, both ozonation and

H2O2/UV systems proved to be effective in inducing diclofenac degradation, ensuring a complete conversion of the

chlorine into chloride ions and degrees of mineralization of 32% for ozonation and 39% for H2O2/UV after a 90min

treatment.

The reactions were found to follow similar, but not identical, reaction pathways leading to hydroxylated

intermediates (e.g. 2-[(2,6-dichlorophenyl)amino]-5-hydroxyphenylacetic acid) and C–N cleavage products (notably 2,5-

dihydroxyphenylacetic acid) through competitive routes. Subsequent oxidative ring cleavage leads to carboxylic acid

fragments via classic degradation pathways. In the pH range 5.0–6.0 kinetic constants (1.76� 104–1.84� 104M�1 s�1)

were estimated for diclofenac ozonation.

r 2003 Elsevier Ltd. All rights reserved.

Keywords: Diclofenac; Advanced oxidation processes; Ozone; UV/H2O2; Kinetics; Drugs

1. Introduction

In recent years pharmaceutical drugs have emerged as

a novel class of water contaminants for which public and

scientific concern is increasing steadily because of the

potential impact on human health and the environment

even at trace levels [1–5]. Huge amounts of these

chemicals in terms of thousands of tons [6–8] are

annually used for therapeutic purposes or in animal

farming in each European country, and may be excreted

both unmetabolized and as active metabolites. Improper

ing author. Tel.: +390817682251; fax:

6.

ess: roberto.andreozzi@unina.it (R. Andreozzi).

e front matter r 2003 Elsevier Ltd. All rights reserve

atres.2003.09.028

disposal or industrial waste may also contribute to their

occurrence in aquatic environments, due in part to

ineffective degradation in sewage treatment plants [9–11].

A typical case is (2-[20,60-dichlorophenyl)amino]phe-

nylacetic acid) (diclofenac, 1), a popular non-steroidal

anti-inflammatory drug widely used to treat inflamma-

tory and painful diseases of rheumatic and non-

rheumatic origin which has been detected in many

municipal sewage treatment plant (STP) effluents

[12,13]. Although it is susceptible to photodegradation

by complex mechanisms depending on environmental

conditions [14], its presence has been documented in

river and lake waters [14–16]. Preliminary investigations

concerning diclofenac impact on aquatic life indicated

some adverse effects on rainbow trouts exposed to water

d.

ARTICLE IN PRESS

Nomenclature

Do3 diffusivity of ozone in water,

1.77� 10�7 dm2 s�1

E enhancement factor, dimensionless

kLo gas–liquid phase mass transfer coefficient

without chemical reaction,

4.26� 10�4 dm s�1

kLoa gas–liquid phase volumetric of mass transfer

coefficient without chemical reaction,

0.045 s�1

ko3 rate constant of diclofenac ozonation,

M�1 s�1

I ionic strength, 0.1M

[O3]B ozone bubbles concentration, M

[O3]F ozone freeboard concentration, M

[O3]in initial ozone concentration, in the gas phase,

1.0� 10�3MQ gas flow rate, 0.01L s�1

rdiclofenac reaction rate, M�1 s�1

VB bubbles volume, 0.034L

VF freeboard volume, 0.2 L

VL volume of solution for ozonation experi-

ments, 0.80L

z stoichiometric coefficient, dimensionless

Greek symbols

a Ostwald coefficient, 0.186 dimensionless

b selectivity degree, dimensionless

s standard deviation, %

D. Vogna et al. / Water Research 38 (2004) 414–422 415

concentrations of 1.0 mgL�1 for 28 days [17]. Although

no conclusive data is available about the possible

environmental effect of diclofenac in surface waters, an

evaluation of the suitability of currently available

processes for removal of organic pollutants from STP

effluents seemed appropriate.

Recently, some investigations on the removal of

diclofenac from contaminated and drinking waters by

means of advanced oxidation processes (O3, O3/H2O2 and

photo–Fenton), have appeared [18–20]; however, neither

detailed kinetic assessments nor intermediate/product

analysis during the degradation pathways were provided.

In the present study, we report the oxidation behavior

of diclofenac to ozone and UV/H2O2, two widely used

oxidizing systems with established effectiveness and a

high level of technical development for industrial

application. Specific aims of the study were to investi-

gate the kinetics of the processes and to assess the nature

of the main intermediates along the route to mineraliza-

tion. This latter aspect is of particular relevance, since

potentially toxic species may be produced by incomplete

degradation of organic pollutants and their assessment

is an absolute requirement should the method be applied

for water remediation.

2. Materials and methods

2.1. Chemicals

Hydrogen peroxide (30% w/w) was purchased from

Fluka. Diclofenac (2-[20,60-dichlorophenyl)amino]phe-

nylacetic acid) (1) was purchased as sodium salt from

Sigma. Glyoxal and oxalic, formic, glyoxalic, ketoma-

lonic, malonic, succinic, fumaric, tartronic, hydroxysuc-

cinic (malic), oxaloacetic, D,L tartaric, 2,5-

dihydroxyphenylacetic (homogentisic acid), 2-hydroxy-

phenylacetic acids and 2,6-dichloroaniline were pur-

chased from Aldrich. All other chemicals were of the

highest grade commercially available and were used as

received. Doubly glass-distilled water was used for the

preparation of standard aqueous solutions.

2.2. Ozonation

A Fischer Model 502 apparatus was used for the

production of ozone. The ozone stream (36Lh�1) was

continuously introduced through a porous sparger in the

semi-batch glass reactor [21] contanining the aqueous

solution (1.090L), thermostated at 25�C. The ozone in

the freeboard phase was evaluated continuously by

means of a Varian UV spectrophotometer at 253 nm

(e03=3200M�1 cm�1) equipped with a quartz cell

(optical length=2.0mm).

For kinetic experiments, ozonation runs were carried

out at three different pH values (5.0, 5.5 and 6.0) in the

presence of a radical scavenger (tert-butyl alcohol,

4.0� 10�3L) to prevent radical oxidation. Phosphate

buffers were used to fix ionic strength at 0.1M, and the

pH of the ozonation mixtures was monitored throughout

the reaction course. Aliquots of the mixtures were with-

drawn periodically and immediately analyzed by HPLC.

Batch ozonation experiments were performed to

check the reliability of kinetic constants. In the case of

ozonation, the experiments were conducted by rapidly

mixing 5.0mL of an aqueous solution, buffered at pH

5.0 and containing tert-butyl alcohol (0.2mL), diclofe-

nac (1.0� 10�4M) and phenol (1.0� 10�4M) with

5.0mL of an ozone-saturated aqueous solution

(1.0� 10�4M) at the same pH. The experiments wererepeated with varying starting ozone concentrations.

2.3. UV/H2O2 oxidation

Irradiation experiments were carried out in an

annular reactor (0.420L) thermostated at 25�C and

ARTICLE IN PRESSD. Vogna et al. / Water Research 38 (2004) 414–422416

equipped with a 17W low-pressure mercury monochro-

matic lamp emitting at 254 nm (Helios Italquartz).

The irradiating power of the lamp (I0=2.7� 10�6 E s�1)

was measured by an actinometric procedure [22].

Hydrogen peroxide concentration during actinometric

runs was determined by a modified iodometric

method [23].

The pH of 1 solutions was adjusted to the desired

values with perchloric acid and/or sodium hydroxide.

The reaction course was monitored by periodical HPLC

analysis.

2.4. Analytical methods

GC-MS analyses were carried out on a Saturn

2000 apparatus (Varian) equipped with an Ion Trap

detector. The flow rate of the carrier gas (Helium)

was set at 1mLmin�1. A DB5-MS fused silica

column (Zorbax, 30m� 0.25mm ID, 0.25 mm film

thickness) was used. The temperature program was

as follows: 80�C for 1min, 7�Cmin�1 up to 150�C,

hold time 5min, 7�Cmin�1 up to 200�C, hold time

5min. The injector and transferline temperatures were

250�C and 170�C respectively. The MS detector

was operated in the EI mode, scanning in the range

40–640 amu. Benzopyrene was used as internal stan-

dard in all runs. Formation yields of reaction pro-

ducts were determined by comparing integrated

peak areas with external calibration curves obtained

for 1.

HPLC measurements were performed on a Hewlett

Packard 1100 liquid chromatograph using a UV diode

array detector. For analysis of 1 a Sinergy RP-max

column was used, eluant CH3CN 0.07M phosphate

buffer pH 2 containing 5% CH3OH, 60:40. The flow

rate was 1.0mLmin�1, and the detection wavelength

was set at 274 nm.

To monitor free chloride formation during the

ozonation and H2O2/UV processes an Orion 96-17B

combinated electrode was used.1H and 13C NMR spectra were recorded at 400.1

and 100.6MHz, respectively. An instrument fitted

with a 5mm 1H/broadband gradient probe with

inverse geometry was used. Experiments used were

standard Bruker implementations of gradient-selected

versions of inverse (1H detected) heteronuclear multi-

ple quantum coherence (HMQC) and heteronuclear

multiple bond correlation (HMBC) experiments. The

HMBC experiments used a 100ms long-range coupling

delay.

Total organic carbon (TOC) was monitored by a TOC

analyzer (Shimadzu 5000 A). The molar extinction

coefficient of diclofenac at 254 nm was determined

under the reaction conditions. The absorbance of the

UV radiation mixtures was determined at different

reaction times.

2.5. Product analysis

Aliquots (1.0mL) of the ozonation mixture, buffered

at pH 7.0, with 1 at 5.0� 10�3 M, were withdrawn atdifferent reaction times, lyophilized (Labconco) and

derivatized according to a reported procedure [24]. After

centrifugation, the mixtures were analyzed by GC-MS.

National Institute of Standards and Technology Library

searching was used for compound identification. Identi-

fication of the most abundant components was secured

by comparison of the chromatographic behavior and

fragmentation patterns with those of authentic samples.

A similar procedure was used for analysis of the UV/

H2O2 oxidation mixture of 1 (1.0� 10�3 M), with

hydrogen peroxide concentration at 0.1 or 1.0M.

2.6. Isolation and characterization of 2-[20,60-

dichlorophenyl)amino]-5-hydroxyphenylacetic acid (2)

A solution of 1 (1.0� 10�3 M) in water (200mL)

adjusted to pH 7.0 was bubbled with ozone for 5min

and then nitrogen to purge ozone. The mixture was

extracted with ethyl acetate (3� 100mL), and the

combined organic phases dried with anhydrous Na2SO4and evaporated to dryness. Fractionation by preparative

thin layer chromatography on precoated silica gel F-254

plates from Merck (0.50mm) allowed isolation of pure 2

as colorless oil (Rf 0.73, CHCl3: CH3OH 9:1, 5mg, 10%

yield); UV lmax (MeOH) 278 nm;1H-NMR and 13C-

NMR (see Table 2).

2.7. Preparation of 2-aminophenylacetic acid (9)

A solution of 2-nitrophenylacetic acid (0.5 g) in 0.2M

carbonate buffer pH 10 (40mL) was treated with solid

Na2S2O4 (4.6 g) at room temperature under stirring.

After 15min, the mixture was extracted with ethyl

acetate (3� 30mL). The combined organic layers wererepeatedly washed with brine and water, dried over

sodium sulfate and taken to dryness to yield compound

2 as a colourless solid (390mg, 93% yield). GC-MS

analysis (O-TMS derivative) Rt=11.45min, m/z 295

(M+).

3. Results and discussion

3.1. Ozonation and H2O2/UV processes

Figs. 1 and 2 show the removal of 1 (1.0� 10�3M) inaqueous solutions buffered at pH=7.0 by ozonation

and H2O2/UV oxidation. Data show complete abate-

ment of 1 within 10min by the ozonation process and

after more than 90min by the UV/H2O2 system. After

90min, the ozonation process induced about 32% of

ARTICLE IN PRESS

Fig. 2. Decay of 1 (1.0� 10�3M), chloride realease and TOCremoved by UV irradiation in the absence or in the presence of

H2O2 (5.0� 10�3M) at pH=7.0: (Full symbols) with H2O2;

(empty symbols) no H2O2 added; 1: circles; chloride: squares;

and TOC removed: diamonds.

Fig. 1. Ozonation of 1 (1.0� 10�3M) at pH=7.0; (�)substrate; (’) chloride; and (E) TOC removed.

D. Vogna et al. / Water Research 38 (2004) 414–422 417

mineralization with a satisfactory chlorine release as

chloride (95%).

In the case of the UV/H2O2 treatment (Fig. 2), direct

photolysis contributes significantly to substrate decay

during the first 30min. Within the same reaction times

the dechlorination rate is not influenced by the addition

of H2O2. After 30min, in the absence of H2O2,

photolytic substrate decay and chloride release were

markedly slowed down, probably because of the

accumulation of reaction intermediates acting as ‘‘inner

filters’’ [25]. Spectrophotometric experiments allowed to

estimate a 43% contribution of the reaction products to

the overall absorbance of the oxidation mixture at

254 nm at 30min.

At 90min of H2O2/UV treatment, the extent of

dechlorination was 51% indicating persistence of sub-

stantial amounts refractory chlorinated intermediates,

whereas TOC abatement was 39%. No mineralization

occurred with UV irradiation alone at this reaction time.

3.2. Intermediate/product analysis and reaction

mechanisms in the ozonation and H2O2/UV processes

The chemical course of the ozonation and UV/H2O2oxidation of 1 (1.0� 10�3M) was investigated by GC-MS, HPLC and TLC analysis. Structural identifications

by GC-MS analysis were supported by comparison with

authentic samples or were based on matching of MS

data. Main aromatic species that could be identified in

the early stages are summarized in Table 1.

Ozonation afforded as chief intermediates 2,5-dihy-

droxyphenylacetic acid (6), homogentisic acid and 2-

[2,6-dichlorophenyl)amino]-5-hydroxyphenylacetic acid

(2). The latter was isolated by preparative TLC and

characterized by ID/2D NMR analysis. Assignment of

the proton and carbon resonances, as deduced by one

and multiple bond 1H, 13C correlation experiments, are

shown in Table 2. At 5min reaction time 2 was 30% of

the reacted material, whereas 6 was about 4%. Other

two components of the reaction mixtures eluted at 23.59

and 23.95min (ca 2% yield each) exhibited a molecular

ion peak (m/z 457) consistent with mono hydroxylated

derivatives of 1. The fragmentation patterns were closely

similar to that of 2 and did not allow to discriminate

among all conceivable isomers. Based on chemical

consideration of the electrophilic character of ozone,

hydroxy derivatives of 1 at 3, and 40 position (structures

3 and 4 in Table 1) appeared more likely. An alternate

hydroxylation product would arise by direct attack of

ozone on the amino group of 1 as proposed in a recent

paper [26]. Careful analysis of product pattern in the

GC-MS trace did not provide evidence for detectable

amounts of nitro compounds, e.g. 2-nitrophenylacetic

acid, which might arise from oxidation at the nitrogen

atom of 1 or of reaction products such as 2-nitrophe-

nylacetic acid, which might arise from oxidation at the

nitrogen atom of 1 or of reaction products such as 2-

aminophenylacetic acid (9). Such reactivity, however,

seems to be peculiar of highly nucleophilic aliphatic

amines [27], but represents a minor route in the case of

aniline derivatives [28].

2,6-Dichlorohydroquinone (10), 2,6-dichloroaniline

(7), 2-hydroxyphenylacetic acid (5) and 2,3-dihydrox-

yphenylacetic acid (8) were also consituents of the

ozonation mixture UV/H2O2 treatment yielded likewise

in addition to 6, 2 and the monohydroxylated deriva-

tives of 1 (3 and 4) in comparable amounts, typically ca

1% yield each at 20min reaction time. Other minor

products of the UV/H2O2 reaction mixtures, which were

lacking in the ozonation mixtures, were 9 and 11, in

ARTICLE IN PRESS

Table 1

Intermediates and corresponding main mass fragments of derivatized by products identified by means of GC-MS analysis in the

ozonation and H2O2/UV oxidation of diclofenac

aShown as TMS derivatives. bStructural formulation can be reversed.

D. Vogna et al. / Water Research 38 (2004) 414–422418

ARTICLE IN PRESS

Table 21H and 13C resonances of 2

Position 1 2 3 4 5 6 10 20 30 40 CH2 COOH

1H(ppm) 6.70 6.55 6.36 7.33 6.95 3.70

(J, Hz) (2.4) (8.4,2.4) (8.4) (8.4) (8.4)13C(ppm) 130.2 119.2 154.9 116.2 122.8 137.2 142.1 130.0 130.9 125.0 40.0 176.7

Scheme 1. Chemical pathways UV/H2O2 oxidation and ozonation of 1.

D. Vogna et al. / Water Research 38 (2004) 414–422 419

which one chlorine atom was replaced by an OH group.

2,6-Dichlorophenol was apparently lacking in both

reaction mixtures.

Non-aromatic fragmentation products that could be

identified by GC-MS and/or HPLC included glyoxal

and oxalic, formic, glyoxalic, ketomalonic, malonic,

succinic, fumaric, tartronic, hydroxysuccinic (malic),

oxaloacetic and tartaric acids.

Based on these results, an overall view of the

advanced oxidation pathways of 1 is depicted in

Scheme 1. Product analysis suggested that UV/H2O2oxidation of 1 proceeds via five competing hydroxyla-

tion routes (a–e). Three of these involve the phenylacetic

acid moiety through the 5-, 3- and 2-positions, leading to

2, 3 and 2-hydroxyphenylacetic acid (5) plus 2,6-

dichloroaniline (7) as primary intermediates (routes a–c),

whereas the fourth stems from hydroxylation of the

dihalogenated unit at C-40, to afford 4 (route d). The

observed pattern of products would suggest that sites

with the highest electron density determine the regio-

chemical outcome of the hydroxylation processes.

Preliminary qualitative inspection of the HOMO coeffi-

cients and total charge density of 1 by semiempirical

methods (AM1/PM3) consistently predicted a higher

reactivity of the phenylacetic unit toward electrophilic

species, with relatively large HOMO coefficients on the

C-5 and C-40 carbons (data not shown).

In this context, formation of 11 would be compatible

with both a direct photolytic mechanism by one chlorine

photosubstitution of 1, as reported in the literature [23]

and OH radical attack at a chlorine-substituted site

(route e). Available data do not allow to distinguish

ARTICLE IN PRESS

Scheme 2. Proposed mechanisms of formation of 2 in the ozonation or UV/H2O2 oxidation of 1.

D. Vogna et al. / Water Research 38 (2004) 414–422420

between these options. It should be noted that direct UV

photolysis of 1, in the absence of H2O2, resulted in about

50% substrate consumption after 90min irradiation

without a defined pattern of products (Fig. 2).

Whereas in the UV/H2O2 oxidation treatment all of

the possible hydroxylation routes seemed to be opera-

tive, the ozonation reaction generated 2 and homo-

gentisic acid (6) as the principal intermediates (routes

a,c), suggesting that hydroxylation at C-5 is the chief

decomposition route of 1 with ozone. The apparent

selectivity associated with ozonation would indicate

substantial mechanistic differences with respect to the

UV/H2O2 oxidation process, in which most of the

degradative paths are mediated by the highly reactive,

relatively unselective OH radicals. Representative hy-

droxylation mechanisms at C-5 leading to 2 are depicted

in Scheme 2 for both processes.

Further hydroxylation of 2 at the activated 2-position

would then induce fission of the C–N bond leading to

2,5-dihydroxyphenylacetic (6) acid and 2,6 dichloroani-

line (7). Similarly, hydroxylation of 3 would trigger an

alternate C–N cleavage route leading to 2,3-dihydrox-

yphenylacetic acid along with 2,6-dichloroaniline (7).

The presence of the OH group on the C–40 in 4 would

direct oxidation chemistry toward the otherwise poorly

reactive dichloroaniline moiety favoring OH radical

attack at the 10-position. Subsequent cleavage of the C–

N bond would then yield 2,6-dichlorohydroquinone (10)

and 2-aminophenylacetic acid (9). This latter product

was demonstrated in small amounts only in the UV/

H2O2 process, suggesting that the relevant route is of

minor importance in the case of the ozonation process.

Failure to detect 2,6-dichlorophenol among main aro-

matic intermediates indicates that of the two NH-

bearing positions, the 10-position on the dichlorinated

ring is not sufficiently activated toward OH radical

attack compared to the 2-position on the phenylacetic

moiety.

Photochemical decomposition of 1 in water was

reported to afford as principal products carbazole-1-

acetic acid (in the absence of H-atom donors), 8-

chlorocarbazole-1-acetic acid and 8-hydroxycarbazole-

1-acetic acid, via dechlorination paths that have been

implicated to account for the behavior and fate of 1 in

aquatic environments [29,30]. It appears, on this basis,

that carbazole derivatives can be taken as specific

indicators of purely photolytic processes. Careful

analysis of the UV/H2O2 oxidation mixtures of 1

revealed only trace amounts of carbazole derivatives

(Table 1, 12), thus suggesting that most of the UV

radiation was absorbed by H2O2 and that direct

photolysis of 1 was not a main contributing reaction

pathway under the specific conditions of this study.

3.3. Kinetic studies

A suitable kinetic model was developed to describe the

consumption of diclofenac in early ozonation stages. It

is assumed that the process develops under a ‘‘fast’’

kinetic regime of absorption with reaction. This model

combines a fluid-dynamic submodel [31] with a simpli-

fied kinetic one in which the ozone attack on diclofenac

is described by one single reaction:

where b parameter indicates the selectivity of the

oxidation for pathway 1.

The complete model includes the differential mass

balance equations for ozone in the bubble ([O3]B),

freeboard ([O3]F) and liquid ([O3]L) phases, each one

considered as a well-mixed stirred reactor:

d½O3�Bdt

¼Q

VBð½O3�in � ½O3�BÞ �

k0La E ð½O3�B a� ½O3�LÞVB

VL;

ð1Þ

ARTICLE IN PRESS

Table 3

Kinetic parameters obtained from experimental runs at

different pH values. [Diclofenac]o=3.0� 10�4M

pH 5.0 5.5 6.0

104 ko3 (M�1 s�1) 1.7670.14 1.6970.10 1.8470.15

% s Diclofenac 8.5 6.3 6.2

% sCl 8.3 6.3 10.5

% so3 freeb 3.4 2.3 4.1

D. Vogna et al. / Water Research 38 (2004) 414–422 421

d½O3�Fdt

¼Q

VFð½O3�B � ½O3�FÞ; ð2Þ

d½O3�Ldt

¼ k0La E ð½O3�B a� ½O3�LÞ � z rDiclofenac ð3Þ

where the enhancement factor E was calculated through

the following formulae:

E ¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1þ

Do3ko3 ½S�z

k02

L

s: ð4Þ

The overall diclofenac consumption and chloride ion

production rates are given by

d½S�dt

¼ �1

zko3 ½S�½O3�L ¼ rDiclofenac; ð5Þ

d½Cl�dt

¼ �bko3 ½S�½O3�L: ð6Þ

For kL0a, kL

0 , a, Do3 ; VB VL and VF the values reported in

the legend (p. 19) were adopted whereas from the data

collected during the runs a selectivity (b) equal to 0.70and a stoichiometric coefficient (z) equal to 2.0 were

calculated.

The model allowed thus to estimate, by means

of an optimizing procedure through which the minimum

of the objective function is determined [32], the best

values of kinetic constant ko3 (Table 3). Statistical

indices such as the percentage standard deviations (s)were used to evaluate the model adequacy. They are

defined as

1

%Y�

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiPðYi � CiÞ

2

N � 1

s

in which Yi and Ci are calculated and experimental

concentrations at the time ti, N and %Y the number of

experimental data and the mean concentration in a

single run, respectively. Data analysis indicated satisfac-

tory accuracy in the estimation of the kinetic constant as

well as low values of percentage standard deviations for

the measured species (substrate, chloride and ozone in

the liquid bulk).

Huber and co-authors [26] reported a kinetic constant

for diclofenac ozonation of B1.0� 106M�1 s�1 deter-

mined by a comparison method with phenol in the pH

range 6.7–7.0. Since this value does not agree with the

kinetic constants estimated above, a new set of batch

experiments were carried out at pH 5.0, by submitting

diclofenac and phenol simultaneously to ozonation. A

kphenol/kdiclofenac ratio of kinetic constants equal to 1.24

was thus calculated from the results of these experi-

ments. By taking into account the values reported by

Hoign"e and Bader [33] for phenol and phenolate ion, a

value for the kinetic constant of ozonation of phenol

(kphenol) equal to 1.89� 104M�1 s�1 was derived at pH

5.0. On the basis of this result and the previously

calculated kinetic constant ratio, a value for kdiclofenacequal to 1.52� 104M�1 s�1 was obtained which well

agrees with those reported in Table 3 and estimated in

the semibatch ozonation runs on diclofenac.

For the H2O2/UV system the determination of the

kinetic constant for OH attack on the substrate by

means of the competition kinetics method [34] was

prevented by the occurrence of the direct photolytic

degradation of diclofenac (Fig. 2).

4. Conclusions

In aqueous solutions under the experimental condi-

tions adopted, both ozonation and H2O2/UV systems

are effective in inducing diclofenac degradation. After

90min of treatment, the ozonation process ensures a

complete conversion of chlorine into chloride ions with a

degree of mineralization of 32% whereas only 52% of

chlorine was converted into chloride ions with H2O2/UV

system with a mineralization of 39%.

The reactions were found to follow similar, but not

identical, reaction pathways leading to hydroxylated

intermediates (e.g. 2-[(2,6-dichlorophenyl)amino]-5-hy-

droxyphenylacetic acid (2)) and C–N cleavage products

(notably 2,5-dihydroxyphenylacetic acid (6, homogenti-

sic acid) through competitive routes. Subsequent oxida-

tive ring cleavage leads to carboxylic acid fragments via

classic degradation pathways.

In the pH range 5.0–6.0 kinetic constants (1.76� 104–1.84� 104M�1 s�1) were estimated for diclofenac ozo-

nation. Limited solubility at lower pH and foaming of

diclofenac aqueous solutions at pH>6.0 prevented the

investigation in a broader range.

Acknowledgements

We wish to thank the Commission of the European

Communities for the financial support of this work

under Grant No. EVK1-CT-2000-00048. We also thank

the ‘‘Centro Interdipartimentale di Metodologie Chimi-

co Fisiche’’ of Naples University for NMR facilities,

Mrs. Silvana Corsani and Dr. Italo Giudicianni for

assistance.

ARTICLE IN PRESSD. Vogna et al. / Water Research 38 (2004) 414–422422

References

[1] Halling-Sorensen B, Nielsen SN, Lanzky PF, Ingerslev F,

Lutzhorft HC, Jorgensen SE. Occurrence, fate and effects

of pharmaceutical substances in the environment—a

review. Chemosphere 1998;36:357–93.

[2] Daughton CG, Ternes TA. Pharmaceuticals and personal

care products in the environment: agents of subtle change?

Environ Health Perspect 1999;107:907–38.

[3] Buser HR, Poiger T, Muller MD. Occurrence and

environmental behavior of the chiral pharmaceutical drug

Ibuprofen in surface waters and in wastewater. Environ Sci

Technol 1999;33:2529–35.

[4] Heberer T. Occurrence, fate and removal of pharmaceu-

tical residues in the aquatic environment: a review of recent

research data. Toxicol Lett 2002;131:5–17.

[5] Bort R, Ponsoda X, Jover R, Gomez-Lechon J, Castell JV.

Diclofenac toxicity to hepatocytes: a role for drug

metabolism in cell toxicity. J Pharmacol Exp Ther 1999;

288:65–72.

[6] Tolls J. Sorption of veterinary pharmaceuticals in soils: a

review. Environ Sci Technol 2001;35(17):3397–406.

[7] Jones QAH, Voulvoulis N, Lester JN. Aquatic environ-

mental assessmet of the top 25 English prescription

pharmaceuticals. Water Res 2002;36:5013–22.

[8] Kummerer K, editor. Pharmaceuticals in the environment:

sources, fate, effects and risks, (1st ed.) Berlin: Springer,

2001.

[9] Holm JH, Rugge K, Bjerg PL, Christensen TH. Occur-

rence and distribution of pharmaceutical organic com-

pounds in the groundwater downgradient of a landfill

(Griendsted, Denmark). Environ Sci Technol 1995;29:

1415–20.

[10] Hirsch R, Ternes T, Haberer K, Kratz KL. Occurrence of

antibiotics in the aquatic environment. Sci Tot Environ

1999;225:109–18.

[11] Zuccato E, Calamari D, Natangelo M, Fanelli R. Presence

of therapeutic drugs in the environment. Lancet 2000;

355:1789–90.

[12] Ternes TA. Occurrence of drugs in German sewage

treatment plants and rivers. Water Res 1998;32:3245–60.

[13] Ternes TA, Bonerz M, Schmidt T. Determination of

neutral pharmaceutical in wastewater and rivers by liquid

chromatography–electrospray tandem mass spectrometry.

J. Chromatogr A 2001;938:175–85.

[14] Buser H, Poiger T, Muller MD. Occurrence and fate of the

pharmaceutical drug diclofenac in surface waters. Environ

Sci Technol 1998;32:3449–56.

[15] Buser HR, Muller MD, Norbert T. Occurrence of the

pharmaceutical drug clofibric acid and the herbicide

mecoprop in various Swiss lakes and in the North Sea.

Environ Sci Technol 1998;32:188–92.

[16] Ahrer W, Scherwenik E, Buchberger W. Determination of

drug residues in water by the combination of liquid

chromatography or capillary electrophoresis with electro-

spray mass spectrometry. J Chromatogr A 2001;919:69–78.

[17] Triebskorn R, Heyd A, Casper H, Kohler Heinz-R,

Ferling H, Negele R, Schwaiger J. Nephrotoxicity of

anti-inflammatory drug diclofenac in rainbow trout

(Oncorhynchus mykiss). Book of Abstracts of SETAC

Europe 12th Annual Meeting: Challenges in Environmen-

tal Risk Assessment and Modeling, Linking basic and

applied research, Vienna, Austria; 2002.

[18] Zwiener C, Frimmel FH. Oxidative treatment of pharma-

ceuticals in water. Water Res 2000;34:1881–5.

[19] Ravina M, Campanella L, Kiwi J. Accelerated mineraliza-

tion of the drug Diclofenac via Fenton reactions in a

concentric Photo-reactor. Water Res 2002;36:3553–60.

[20] Ternes TA, Meisenheimer M, Mcdowell D, Sacher F,

Brauch HJ, Haist-Glude B, Press G, Wilme U, Zulei-

Seibert N. Removal of pharmaceuticals during dringing

water treatment. Environ Sci Technol 2002;36:3855–63.

[21] Caprio V, Insola A. Aniline and anilinium ion ozonation

in aqueous solution. Ozone Sci Eng 1985;7:169–73.

[22] Nicole I, De Laat J, Dor"e M, Duguet JP, Bonnel C. Use of

UV radiation in water treatment: measurement of photo-

nic flux by hydrogen peroxide actinometry. Water Res

1990;24:157–68.

[23] Treadwell FP, Hall J. Analytical chemistry, vol. II, 8th ed.

New York: Wiley; 1935.

[24] Li X, Cubbage JW, Jenks WS. Photocatalytic degradation

of 4-chlorophenol. 2. The 4-chlorocatechol pathway. J Org

Chem 1999;64:8525–36.

[25] Beltran FJ, Gonzales M, Gonzales F. Industrial waste-

water advanced oxidation. Part 1. UV radiation in the

presence and absence of hydrogen peroxide. Water Res

1997;31:2405–14.

[26] Huber MM, Canonica S, Park GY, Von Gunten U.

Oxidation of pharmaceuticals during ozonation and

advanced oxidation processes. Environ Sci Technol 2003;

37:1016–24.

[27] Munoz F, von Sonntag C. The reactions of ozone with

tertiary amines including the complexing agents nitrilo-

triacetic acid (NTA) and ethylenediaminetetraacetic acid

(EDTA) in aqueous solution. J Chem Soc Perkin Trans

2000;2:2029–33.

[28] Sarasa J, Cort"es S, Ormad P, Gracia R, Ovelleiro JL.

Study of the aromatic by-products formed from ozonation

of anilines in aqueous solution. Water Res 2002;17:185–94.

[29] Poiger T, Buser HR, Muller DM. Photodegradation of the

pharmaceutical drug diclofenac in a lake: pathway, field

measurements and mathematical modeling. Environ Sci

Technol 2001;20:256–63.

[30] Moore BE, Thomson SR, Zhen D, Duke CC. Photo-

chemical studies of anti-inflammatory drug diclofenac.

Photochem Photobiol 1990;52:685–90.

[31] Andreozzi R, Caprio V, Insola A, Tufano V. Measuring

ozonation rate constants in gas–liquid reaction under the

kinetic-diffusional transition regime. Chem Eng Commun

1996;143:195–7.

[32] Reklaitis GV, Ravindran A, Regsdell KM. Engineering

optimization. New York: Wiley; 1983.

[33] Hoign!e J, Bader H. Rate constants of reactions of ozone

with organic and inorganic compounds in water-II.

Dissociating organic compounds. Water Res 1983;17:

185–94.

[34] Onstein P, Stefan MI, Bolton JR. Competition kinetics

method for the determination of rate constants for the

reaction of hydroxyl radicals with organic pollutants using

UV/H2O2 advanced oxidation technology: the rate con-

stants for the tert-butyl formate esters and 2,4-dinitrophe-

nol. J Adv Oxidation Technol 1999;4(2):231–6.